roland thinks that brains-N-brawn might be too obnoxious of a domain name. and that might be a reason i worked so little this last year. he's definitely right about the domain name being obnoxious. that was intentional. and mainly as a self challenge. if i was going to have the balls to brand myself that way ... then i better work my ass off to put some meat behind it. i've certainly tried to meet that challenge. ashamedly, i have let the brawn aspect slip ... and am in the process of beginning a new workout program. but i question that the name has kept potential employers at bay? 95% of my contracts have been through recruiters. when presenting me to a client, they typically take my resume and stuff it into their format. in that phase they usually strip off all of my contact info. e.g. the articles i write are most likely not seen by a client. from experience, i've never been to a client that actually knew about my website before i showed it to them. yes, i find that a little disheartening. on the other hand, i have been to clients that cannot access my website because the firewalls have it blacklisted as being adult content. that is a different story ... and would be a compelling reason to change domain names. but my guess is that bra-N-brew would get blacklisted as well :) so from a marketing standpoint, this website has been a complete and total failure. from a self-motivating view ... its a success

mark your calendars [info]

i dont do resolutions, per se, but I do come up with goals at the start of each year. had 3 primary goals for this past year ...

1) blog (+). it was mainly supposed to be a moblog ... which is implemented, but i dont use (-). unexpected happenings include getting linked by some A listers (+) and not getting at least 1 job offer because of said blog (-) ... actually that is probably a good thing in the long run (+)
2) MCSD .NET (+). that was the one check box i did not have filled whenever a recruiter called. the messed up part is i only worked 3 months last year (-). i'm not exactly sure why my skill set is not in demand? oh well, at least my taxes will be low (+) and i had a lot of time to upskill (+). overall, getting certified seems to have been a waste of time (-)
3) Artificial Intelligence (+/-). read ALOT of AI books (+) and wrote a handful of articles (+). yet i've still got ALOT more to learn about AI (-). and i let my AI push get interrupted (-) to start teaching myself Spanish (+).

overall, i give myself a 70% for last year

aside. sometime this week marks my 4th year with .NET. it all started when i installed a VS.NET Alpha because i was snowed in one weekend in NY. turned in my 2 week notice the next work day

... its just a matter of time. not sorry for being cold blooded, but i'm finding the tsunami videos to be altogether unimpressive. its hollywoods fault for desensitizing me. now the body count, that is too impressive. my guess is that they are currently high, much like 9/11 initially was. regardless, this really sucked. and being xmas just makes it tragic. my guess is religous nuts are now tracing it back to passages in the bible or a nostradamus prediction. either way, its kind of appropriate that we've been eyeing MN4 at the same time. because an asteroid could definitely be considered a WMD. thats more along the lines of where i think our resources should be. i'm talking back to basics. e.g. survival. i think most people believe in survival of the fittest, they just dont practice it. in the US, we've made things way too safe. its way too hard for stupid people to die off. and its mainly the legal system thats keeping them alive. but if we've got a comet heading our way ... what good is our legal system going to do?

both times i've been to Harvard, i've heard the story about how students cant graduate until they learn to swim. some people consider this a ridiculous requirement. do you?

all i know is that a cat was cloned for 50K for somebody in TX, and now that company is setting up in Madison, WI. 1st, i think its appropriate that it was for somebody in Bush's state. 2nd, i'm glad their coming to my current state of WI. 3rd, for 50K, that better be one special cat. because cats are FREE from animal shelters. before you know it we will have to get our cats spayed / neutered / and some term to represent a process that will keep a cat from being cloned. 4th, i'm really surprised a dog wasn't cloned before a cat. 5th human cloning is inevitable. based on the top googles searches of last year, we will soon have a cloned society made up of blondes.

which of these happens 1st (if at all)?
a) cloning a human
b) stopping the aging process
c) robot AI killing a human


my bet is they all happen, in that order. and all of them happen before i kick it

i don't 'get' holidays ...

anyway, i predicted the voice browser wars years ago (granted, i did not predict that MS would screw up their 95+% lead with IE). but Opera just fired a warning shot with their new beta. on the IBM / Opera side is X+V, with the V standing for VoiceXml. on the Intel / MS side, the technology is SALT. so for web developers, supporting both these specs is going to be a PITA. the messed up thing is that MS has had a SALT add-in for IE (and Pocket IE) for almost a year. they just have not pushed it out. that is why multimodal Speech SDK apps get no love from developers. hopefully IE .Next will get it right

another holiday i will not be recognizing ... Be Nice to a Recruiter Month, when the rest of the time is about Shafting the Programmer

so what would it take for me to be nice to a recruiter?
1) tell me your profit margin.
2) let me see the contract you sign with the company.
3) dont falsely sign me up for a 6 month contract, and then you end up renewing after 3.
4) when you tell me you will submit my resume ... do it.
5) when you take me out to lunch on my money ... make it steak.


... i could go on. if you want my respect ... you have to earn it. but my personal experience with recruiters has caused nothing but loathing. my personal goal is to get enough name recognition that people will contact me for work directly, and the recruiter is cut out of the picture. i can make more, and the company pays less. win win. granted ... i don't seem to be even close to making that come true :(

was boarding the plane and my carry-on started making a loud buzzing noise. had to wait for people to get seated before i could get to my seat and rummage through my bag to turn whatever it was off. everybody i walked past was looking around to see what was going on. one girl turns to me and says "that better be a razor". then the fight club scene hit me ... "Nine time out of ten, it's an electric razor. But, every once in a while" ... it's an electric toothbrush

my Tablet might be dead ...? was happily minding my own business and teaching myself Spanish on the plane. and them some Scoble scene breaks out. the girl seated across from me asks "what's that?". so i give a little demo and let her write on it and such. then the person seated in front of her turns around and starts checking it out too. finally i get it back ... and then the stewardess wants a demo. i agree ... pending i got an extra cookie. with that over, i go back to Espanol. then i randomly get some idea for something to code in .NET and fire open Journal to ink some UML. go to save ... and nothing. total lockdown. wait 5 minutes ... nothing. while waiting, i make sure the Tablet is nowhere near my genitals so that it cannot damage my fertility :) that just wouldn't be fitting with a last name like chesnut. finally, i power down the machine, and bring it back up ... to this :

STOP: C0000218 Unknown Hard Error
Unknown Hard Error
Beginning dump of physical memory

... that dump of physical memory never ends. been screwing with BIOS and such and haven't gotten any further. i'll have to retrieve the docking station from my parents before i can be sure if it can be salvaged. at least it made it 1.5 years, which says alot, because i'm not exactly easy on mobile devices ... or anything for that matter. this is a very sad day in casey-land

its a shock to me as well. it happened like this : [flashback] last year, Software Legend (the URL doesnt work anymore!) David Platt posted to the WSE newsgroup looking for CF WSE bits, so he could create an incredibly difficult homework assignment for his Harvard .NET class. [what other blog is going to move from Hooters .NET to Harvard .NET?] he asked that question about a day after i had finished the CF WSE implementation, so i responded. anyway i got an invite to present at his class ... and it was a lot of fun. [flashforward] somehow, i even got invited back this year. and it was equally as much fun. as David had warned me, they are one of the most technical groups I have ever presented to. that makes things interesting, because you can assume ALOT of background knowledge on their part. definitely a high quality experience. here are some pics

now for the inbreeding. a former Platt student (or TA?) is going to work for the Tablet group. while we're eating i get a call from somebody on the Tablet team. i come back to the table and ask her to guess who called ... and she guesses correctly. spooky. also, at that dinner table is the TA for the class Jason Haley, who happens to write one of the better blogs that i read. too cool

[ink recognition] B Civil Disobedience? rpt*

isf

i was going through the RosettaStone Spanish lessons ... and found my 1st bug in the picture below. that is definitely an 'el hombre' and not a 'la mujer'. other than that small error ... RosettaStone kicks ass. the one thing that sucks about it is i cant figure out how to get translated sentences within the program? e.g. if i had no clue what that sentence in the picture meant, i would like to hit a key and see it in english. for most pictures you can tell what is going on, but there are some that are questionable. for those i have to open up a PDF to get the english version ... why in the world isnt that just build into the program?

his blog post is here : Getting Web Services using HTTP 1.1 Compression to work in Compact Framework. it would be really interesting to get this synched up with the WS-Compression implementation for WSE 2.0 at Plumbwork Orange.

now for some rattling on about WS speed in CF in general. its common knowledge that they are slow. they are slow for multiple reasons : small processors, framework, and small network pipes. 2 of these are out of our control : hardware and bandwidth. luckily those both keep getting faster. for compression, it leverages extra CPU to make it smaller on the wire. that leaves the framework. on the desktop, if you need speed, you can do remoting. we dont have (nor are we getting) that option for CF, nor do i think we should. granted, it might be possible to port it over from Mono? for WS on the desktop, it gets speed by building an assembly for de/serlialization of an object by reflecting against its object graph. this happens the 1st time the WS is called. every time its called from then on it is faster because it does not have to use reflection to figure out the object graph. we dont get to build an assembly like that for CF on the device ... thus its slow in comparison. er, um ... actually you can build an assembly on the actual device at runtime using Pocket C#. but there is another way. for ASP.NET 2.0, it introduced the concept of SGEN. SGEN does that reflection of the object graph to build a custom serialization assembly when the project is built. so every call is fast, even the 1st one. CF should get to leverage the SGEN capability ... and use it by default. that along with the performance improvements for XML in CF v2 might make WS more attractive. once we get past that performance hurdle, then we can proceed to slow them back down by adding security. then when we get past that, the next step would be to make hosting WS on CF a more viable option.

i lead a double life as a secret agent. below is a collage of my better costumes. these pictures were drawn on this ink-enabled web page from the /tabletWeb article.

actually considered a nose ring when i was a bad ass. anyone else think the 'fat zorro' one makes me look more like the hamburglar?

yes, i'm an idiot ... i couldn't follow the plot. stuff happened ... and i dont know how or why

looks like Sean Campbell is getting into this stuff too. aka 'seamless computing' per Bill G. this is how i've been positioning myself over the last couple years, because it requires having an understanding of a bunch of technologies : speech / ink, mobility, location, web services. obviously security for privacy and other reasons. i've also taken the liberty and added artificial intelligence to the list so that the technology does not become 'obtrusive computing'

will never use MSN spaces because of their censorship crap. now i'm pissed that google suggest is censored as well. lame. my SafeSearch is set to No filtering, so now we need a SafeFilter option to turn that off as well. i'm really pissed that most of my english-spanish software programs have left out dirty words as well. you expect me to go to another country and not cuss or understand when somebody else is cussing ...

the title is a span-glish word-play off of 'to be, or not to be', since both words can mean 'to be'. that's been giving me some problems. other difficulties are verb conjugation (especially irregular verbs), use of double negatives, and adjusting words to gender. the good news is that i've gotten through alot of beginner material and am starting to build up some vocabulary. i was actually in McDonalds last week and they were advertizing their new chicken sandwiches. they had a flier that was english on one side and spanish on the other. was able to read and understand at least 80% of the flier. granted, i then attempted to translate my bottled water label to spanish, and could only convert about 5%.

and out of pure luck i stumbled upon a language expert, who happens to know a smackload of .NET. he's been able to answer many of the questions that i've had about languages in general. one shocking piece of info he pointed out was that English is one of the most difficult languages. that's cool, because i've already learned it. er, um ... at least you've been able to read this far. also, i can reason that the other languages should be easier in comparison. on the other hand, maybe that is why we have so many illiterate people in our country and all the offshore projects i've seen have been crap. e.g. there is no way (at this point in time) i could attempt to develop an application for a spanish employer. reverse that ... i am now better prepared to develop an app for primarily english users that needed to support spanish users as well.

back to learning spanish. the approach to learning a little a day is working great. i've been doing about an hour of day of audio, and another hour of interactive training programs. eventually, i need to work in reading exercises too. dare i say, there has been some sort of 'synergy' using this method. if i miss something the 1st time with audio, i've been able to pick it up through software ... and vice versa. and related to AI, my brains has been working stuff out for me during sleep. i'm having numerous instances of feeling really uncomfortable with a topic before bed, and then the next lesson on the following day i'm able to magically recall words i could not the day before. so my neural net is in serious overdrive. my headaches have mostly gone away too. its as if the barriers to learning a new language have been significantly torn down.

finally, i've been trying to relate this to learning programming languages. human languages definitely have similar classifications such as romansque and intonal vs OO and declarative programming languages. once learning OO, its simple to learn any other OO language. that is why i think spanish has been going so well, since both english and spanish are so related (e.g. moving from Java to C#). if i were trying to learn chinese (intonal) instead of spanish ... that would be like moving from C# to XSLT. the main difference between natural and machine languages is the amount of keywords. in programming, i only have to learn so many. e.g. how do i loop, do comparisons, etc... there is just not that much. but with natural language, i have to learn a large vocabulary to do anything meaningful. the good news is that the subset of vocabulary for common speech is quite small, and increases when writing, and then reading.

aside, my favorite spanish word is 'maybe' ... it sounds just like 'kiss ass'

we had about 10 people show up for the party, and it was a blast. i've uploaded the pics that i took. there were a couple other camera phones there, so i will post those pics as they become available

just a reminder that the 1st Milwaukee area geek dinner will take place this Teusday at the Greenfield Hooters starting at 6 PM (more info). if you have not already ... please contact me if you plan on attending. i've had an disproportionate amount of people contact me wishing they could come :)

i'm not a holiday guy (BAH HUMBUG) ... but i'm putting together an xmas mp3 CD for my car. my tunes consist mostly of techno remixes and comedy. but i just found this! its 1994, but the Frosty the Snowman track is old school excellence (Amazon has a preview). the only thing about bassing during xmas time in Milwaukee is you have to keep your car windows closed, so it does not sound as good as when your windows are open. also the car rattles more because the metal shrinks a little

over the last couple weeks my hotmail account went from the expected 250 megs, then to 2 gigs :), then to 2 megs :(. this week somebody from MSN contacted me, who had been contacted by Scoble ... and now i'm all fixed up. Thanks Robert and MSN person. so the moral of the story is ... if you bitch, then they will fix it. now back to my regulary scheduled bitching about not being able to program PMCs or SPOT watches ...

.netcpu ... per Mike Hall

this is announcing the 1st blogger dinner for the Milwaukee, WI area. and its really going to be more like a party! i'm talking wings, beer, girls in skimpy clothing, and a bunch of smart people talking about .NET and blogging. serious ... last year was Hooters 20th anniversary. you could get a 'passport', and if you visited 20 different Hooters restaurants to get your passport stamped within that year, then you earned a wing party. needless to say ... i completed it easily ... and now i need to have a party! the wings are free, we just have to pay for the beer

it is going to be Teusday Dec. 7th from 6 to 8 PM at the Hooters in Greenfield, WI. you need to email if you can attend. there is a max head count so i can only claim so many in my party. if there is too many people that want to attend, then you can still show up, i just cant get you the free wings. and bring your camera phones :) aside, Hooters spams me periodically ... and its the only spam that i recommend. especially from HootersTexas.com. speaking of that ... i always flinch a little when typing in Hooters.com ... but it does point to the restaurants site, and not pr0n as one might expect

Richard Sprague points to an interesting article about Designing speech apps for Spanish. that is actually a secondary reason for my learning Spanish. if you're going to do a voice-only app in the US ... you basically have to support Spanish too. yes, sometimes there is method to the madness

got to spend this Thanksgiving at a really small town. entertained myself for a couple hours at the local library. this was interesting since i had not been in a library after discovering the internet. the good news is that they had a handful of public computers. there was no filtering software at all ... so i could even browse to my own site (my site is blocked by alot of filtering software ... oh well). the bad news is they did not have any sort of software installed to keep somebody from running spyware or key loggers; nor did they refresh the system after being used by somebody. they also had a good number of VHS and DVDs to checkout, along with music CDs and audio books. the really bad news was the selection of books. their science section consisted of about 35 books. they had no math or computer related books. on the other hand, they had over 100 religious related books. yes ... this was a red state

remember how i had 2 gigs right before Thanksgiving? well somebody fixed that error with another error. they knocked me back down to 2 megs. for a while i had been sitting fat and happy at 250 megs. somebody tell somebody over there to get a clue

er, um ... i had 250 megs last week ... and 2 gigs this week! i was paying for MSN at the start of the year, but cancelled it about 3 months ago. is this a screw up or a pissing contest with gmail? enough already guys ... i dont want that much email ... or space. when i open up the inbox and see that i've only used 0%, it makes me feel kind of lame :( felt kind of cool when i was always close to that 2 meg limit :) those nigerians really like me!

per Hervey ... WSE 2.0 Service Pack 2 PreRelease

this update will break one of the sample calls for CF WSE2 ... specifically WS-SecureConversation. it will break because WSE2 SP2 requires UsernameTokens of RST requests to be encrypted. my CF client code only signs the UsernameToken and does not encrypt it. the CF WSE2 libraries do have the capabilities to encrypt UsernameTokens; look for some commented out code in the UsernameSigning client side code. granted ... i only tested that code with a modified UsernameSigning sample. i did not test it with the SP2 SecureConversation sample

and the real question is ... why do we need WS-SecureConversation for mobile devices? the answer is to reduce the amount of asymmetric cryptography, which is slow and computationally expensive (especially on small devices). WS-SecureConversation allows you to exchange a session key using asymmetric key encryption in the 1st couple calls. for all subsequent calls you can use that key (or derived keys) and only do symmetric encryption (and signing). otherwise, if you were just using the lower level stack of WS-Security, then you would have to do asymmetric crypto in every WS call

i'm relatively healthy and might get a cold (at worst) once a year. used to get a headache maybe once a year as well ... except they are now on the upswing. this has just happened the last couple weeks. believe it or not, it seems to be related to when i'm studying Spanish. i've had this theory for the last couple days and it seems to be holding true. guessing that my neural net is getting a new topology? the only other time there is a high probability of my getting a headache is the 1st day on a contract. when the client turns on the firehose and starts filling my brain with info about a new domain ... that sometimes sets one off. the good news is that the headaches seem to be getting less severe. hopefully, they go away entirely. wonder if they will come back if i attempt a 3rd language? just so long as they dont turn into this train wreck

been reading about Genetic Algorithms (GA). understand the concepts and have started to come up with some article ideas. i've got a couple ideas half way thought out, that might be cool, and that should be feasible. my current view of GA is as a search algorithm. if there is a large problem space, then GA is a quick way to find an answer. also like how it has been applied to neural networks a handful of different ways

this has been tough with AI in general. understanding an AI technique itself is pretty simple. coming up with a cool problem for it to solve is much more difficult. my hope is that once i've got a good sampling of AI techniques in my tool belt, then i can just pick which technique to use depending on what problem i need to solve. the bad news though is that there seems to be a lack of companies trying to solve hard problems. there was a blog post yesterday about somebody being bored with working with data and web apps. agreed ... that is some boring crap. mobility makes stuff more interesting and has more chances to work in AI. games definitely have a need for AI. and robotics cannot happen without AI. so in closing ... web apps pretty much suck

i've listened to the 1st CD of 4 different programs. here is how i rank them from best to worst :

1) Pimsleur - the audio program works with repitition and slowly introducing more words. an English speaker will introduce the word, and then a man will say it in Spanish. each section usually centers around a conversation between a Spanish man and woman. it does a great job of making you recall words that you learned previously and the learning pace is almost perfect for me. after each lesson, i feel like i understood more than 90% of the content. there are 3 programs from beginner to advanced. each program has about 15 hours worth of content, along with a short reading program. the only negative is it is ridiculously expensive. dont know if its a scam, but you might check out cheappimsleur.com
2) Michel Thomas - this is rather unique. you learn by listening to Michel Thomas teach a male and female student how to speak Spanish. so you get to hear what they do wrong ... so you dont make the same mistakes! Michel corrects them and makes them repeat it if they get it wrong. this style works really well. the program i've got has about 8 hours of content. the only problem is that the female student learns faster than me ... she must be cheating! the pace was a little too fast, so i only felt like i understood 75%. also, Michel is sometimes hard to understand
3) EspanaViva - this course has about 3 hours worth of content. it goes through some basic speech. there is a male and female narrator. you also get to listen to some recorded conversations with other speakers. it went too fast and i only understood about 66%
4) Learn Spanish In Your Car - so far this is just a vocabulary course. a girl will say a word in English and then a guy will say it in Spanish. it is monotonous and boring, nor does it ever make you recall previous words. there are 3 hours in total ... so hopefully it changes! maybe i should play the tapes while i'm sleeping and learn though osmosis? only undertood about 50%

out of those, i recommend the 1st two programs. Pimsleur is by far the best to listen to in your car because its recorded quality is really good. Michel Thomas is less clear and the pace is faster, so i would not use it while driving

cannot tell you how many times i've applied

another Tablet guy losing his way. looks like Peter is considering a 'cute little touch screen sub notebook' and a 'full featured Media Center notebook'

from my perspective, i would never buy anything that is described with the adjective 'cute'. sub notebooks are exactly that. if i carried a purse, then i would get one and that is here it would go. but i dont sit to pee, so that is out of the question. have tried to imagine a scenario in which a sub notebook would provide a better experience for anything that i do ... and i'm coming up empty. granted i've never used one ... so maybe i'm wrong? as far as the full featured Media Center notebook ... i do want one of those ... for no apparent reason? i dont watch TV, dont listen to music much, nor do i take a smack load of pictures. there is absolutely no reason for me to want one ... but i do ...

just saw that the 3rd Aqua Teen Hunger Force DVD has been released, with the 'Broodwich' episode! if i watched TV ... this would be the extent of it. more info at tv tome

MY NAME IS ...
Shake-zula, the mic rulah, the old schoolah.
You wanna trip, I'll bring it to ya!
(Yeah) Frylock and I'm on top, rock you like a cop.
Meatwad you're up next, with your knock-knock.
Meatwad make the money see. Meatwad get the honeys G.
Drivin in my car, livin' like a star.
Ice on my fingers and my toes and I'm a Taurus
(Dancing is stupid) Uh, check-check it, yeah
'Cuz we are tha Aqua Teenz,
make the homeys say ho 'n the girlies wanna scream!
'Cuz we are tha Aqua Teenz,
make the homeys say ho 'n the girlies wanna scream!
(Yeah) Aqua Teen Hunger Force. Numba one in tha hood, G.

saw a couple of links to a new beta scholar.google.com (Stand on the shoulders of giants). its targeted at searching research publications. the searches i've tried so far have mainly returned links to PDF and PS files, as well as Books to be followed up by 'library searches'. it also has 'cited by' links to find out what research papers referred to it

scholar.google is a direct knock off of CiteSeer.ist (Scientific Literature Digital Library). CiteSeer is an existing research focused search site that i've been using for the last couple months to search for AI related papers. google adds some additional functionality such as 'view as html', 'web searches' and 'library search'. if you go a level deeper on CiteSeer, it has a summary page for its entries giving more detail plus an abstract. it is worth noting that MS Research is a CiteSeer sponsor ...

on a somewhat related note ... have you ever checked out the Math / Science sections at bookstores? they are absolutely pitiful. Computers get about 5 stacks, and Math / Science is lucky to get 1. this sucks, because i rarely read Computer books anymore ... because its constantly the same crap. do i really need to know how to build a better DataGrid? and all the cuts that VS 2005 has taken ... its more like a point release than a new version. guess that its natural for an industry to slow down over time ... 'simulated annealing' if you will

yes, language wars are stupid ... but for some unknown reason i dont like VB. think it is from doing VBScript back in the ASP days? now i'm having to convert some VB6 code to VB.NET. seems like the wrong time to be adding VB.NET to my resume, especially after the announcement that VB.NET would get minimal refactoring support in VS 2005. actually seems like the right time for VB'ers to jump to C# since it now has edit-and-continue ... which C#'ers mostly dont care for. what is interesting is that i'm actually getting to test a theory. when recruiters call specifically for VB.NET, i've told them that i've done C# for years but could switch at a drop of a hat. of course they dont believe me (the distrust is mutual). regardless, here is my 1st day with VB.NET

[beginning of day]
ends up that the syntax change is as simple as expected. arcane VB6 syntax is slowing me down substantially moreso than the VB.NET syntax. the parts that are odd about VB.NET is declaring variables. for quick code in C# i'll write something like XmlDocument xd; so 'xd' is from the PascalCase of the Type, and it reads in a forward manner. in VB.NET, it wants the Instance name 1st ... Dim xd as XmlDocument. its like thinking forward, then going back, and then forwards again. that misstep is really bugging me because i am not trained to think that way. the other annoyance is muscle memory keeps typing a semicolon at the end of every line; cant stop it; also cannot stop myself from caring about case. as a language it is more verbose and it uses parentheses too much. finally, the IDE tries to be too smart. as i type, it keeps adding code or formatting that i do not want to be added. very annoying

[end of day]
got past most of the annoyances from this morning. e.g. no more semicolons. the IDE is still doing more than what i want it to, but i've figured out how to bypass some of its tricks. so now i'm able to use the language productively ... but there is still something about it that just doesn't feel/look right. hypothetically typing, if there ever was a 'language war' my dog tags would read KC#. slash-and-burn tactics would be in effect and encouraged

remember when the US was the brain drain? now we've got less foreign students entering our universities for advanced degrees. actually, we've got less of our own students obtaining advanced degrees as well ...

as for role reversal, i've been contacted by 2 overseas companies about leaving the US to join them. one of those companies is small, the other large. one of the countries is small, the other large. i've met one of the persons before, the other is a stranger that knows my website. if you're wondering ... neither of these places speak Spanish.

another possible trend. i've had 2 (US based) contract calls to fix projects that were originally outsourced offshore. now as the end of the year approaches, they have deadlines to get the projects completed and touched up. both of them are way over budget and schedule. what happened to people contacting me to work on the latest and greatest stuff?

regardless, i'm just a piss ant, so only my mom would care if i left. but what if Kurt Cagle took his brain to Canada? anyway, the TS has me staying put. you should try convincing us to leave during the dead of winter :)

IsYourJobGoingOffshore.com is pointing to a Forbes article asking : Must the Whole World Speak English?. the article was sparked by some French study ... which isn't important. the point is the population of India will pass China around 2050. as India becomes more fluent in English, the author suspects that China will have to adopt the English language as well.

2050 is some time off. but right now, a friend of mine from Dallas has told me what a telecommunications company is currently doing. they are laying off Managers and Engineers that do not know (or aren't learning) Mandarin Chinese ...

even more interesting is that there was study earlier this year showing that : English is in Decline as a First Language. assuming that is just because our population is not growing as rapidly as others?

to put some things in perspective, here are some numbers i found regarding native speakers. forget where i found these, nor do i know how accurate / recent they are?

1. Mandarin Chinese 836,000,000
2. Hindi 333,000,000
3. Spanish 332,000,000
4. English 322,000,000
5. Bengali 189,000,000
6. Arabic 186,000,000
7. Russian 170,000,000
8. Portuguese 170,000,000
9. Japanese 125,000,000
10. German 98,000,000

really like the concept of Esperanto. had similar ideas for an engineered language as soon as i started looking at Spanish and seeing its oddities. just cannot see myself learning what seems to be a mostly unused language. the studies showing that it helps you pick up a 3rd language much faster is interesting; but i wonder how much that is based on Esperanto itself vs training yourself to learn 'how to learn' another language.

its been interesting to look at the foreign language materials available at HalfPriceBooks. its a decent measure to see the trends that people have pursued in the past. Spanish has the most materials by far ... as expected. after that ... French, Russian, German, and Japanese seem to be equally represented. French and German were expected because i believe those were the other foreign language electives that my high school provided (other than Spanish). the amount of Russian and Japanese materials was unexpected. finally, there were minimal materials for a number of other miscellaneous languages. what shocked me was the lack of materials for Chinese. Arabic and Hindi was mostly non-existent as well

(Mitchell, 1998) this was a great introductory book. it is a quick read and has very little math. the 1st chapter gives you the basics when describing some examples of how GA has been used in the past. chapter 2 gives some more examples of GA and takes it slightly further just into the real of Genetic Programming. 3 steers more toward Artificial Life and to how GA might be used to look at our own world. 4 goes into more detail and 5 explains some of the decisions to be made when implementing a GA. the final chapter is just a couple pages about the future of GA.

Bill reminded me about aftermarket car products. this is another one i want ... i have an older sports car. the problem i have is when an SUV or big truck pulls up right behind me at night. if they've got the newer lights (super bright with a blue tint), then it floods my cab with light and blinds me. so i want a deterrent. there should be a video camera facing out of my back window. if it determines that it is night and a vehicle has obnoxious lights, then i want it to do face recognition in the general vicinity of where the other driver should be. the best solution would involve some way to reflect the bright light from their vehicle off of my car and back into their eyes. maybe some special glass or paint with nano technology. or a mirror. then that would be directing their own annoyance factor against themselves. a simpler approach would be just to have a motorized spotlight in my car too. it would just adjust for parallax between it and the camera, and shine directly in their face to let them know i'm pissed. not sure if you would or wouldn't want this to work when a cop pulls you over?

NOTE that all my aftermarket car ideas involve loud radios or getting revenge on other drivers.

Scoble is asking about How to excite 10-to-13-year-olds about programming and math/science? i think that is the wrong question. the question should be ... How do you get kids interested about programming before they hit puberty? and as the TopCoder says ... the answer is games. violence is certainly what got me started with computers. then puberty hit ... and sex took over. then you just have to tell the kids dub dub dub dot playboy dot com, and you're done. they will spend every moment they can trying to get past whatever content blocking software you have installed. think of it as a rite of passage, such as the spartans leaving their kids on hillsides to survive on their own and find their way back home. the next stage is money. but as the money goes out of our industry, you can always revert back to sex as a motivating factor. another good question is 'How do you get girls into programming?' ... i have no clue ...

could there be more soap operas on the spanish channels? i dont mind all the senioritas, but the overacting is killing me. haven't these people heard of reality TV? based on a comment to this article : How to Learn A Language, will probably try to watch some DVDs with a Spanish audio track instead. also be sure to read the comment directly below the DVD one ... the start of the 1st sentence can be taken entirely out of context :) the article reccommends Pimsleur audio courses. these are too expensive to get from audible.com, but i've listened to them elsewhere and they do sound excellent. for software i am trying 'Instant Immersion', which the article advises against. i'm on my 3rd CD now and it is actually kind of working. the 1st CD 'Talk Now' was easy and all i had to do was memorize vocabulary. the 2nd CD 'World Talk' is much MUCH harder. it is all Spanish (i.e. Immersion) and i basically only got through the exercises by listening for a few key words; but there was alot more happening i did not understand. this bothered me, but being able to hear the keywords is actually the point of the lesson. the 3rd CD seems to be mostly about pronounciation and the 4th is an interactive dictionary. also picked up the 'Learn Now' program, but have not tried it yet. i consider 'Rosetta Stone' to be too expensive and will try to find it cheaper. the Flashcard program for my PPC is 'Spanish Before You Know It', and i consider it excellent. it has a study mode, and then can quiz you based on Enlish-Spanish and Spanish-English. it also has audio to hear the Spanish words being pronounced. will have to try another from the article as well. have not looked at a Spanish Reader ... yet

didn't this blog used to be about technology? going to have to work harder on making the language studies secondary.

like that the alphabet is basically the same as English. like that all words follow simple rules of pronunciation. granted it will be hard for me to break my old habits. really dont like that nouns have gender. who came up with that stupid idea? in French it wasn't so bad because both 'le' (M) and 'la' (F) started with 'L', so you could kind of mumble and ignore gender. 'el' and 'la' will be harder to fake. the rolling R is tough too. i can only manage it at the beginning or words, and do it way too exaggerated. and why do all languages have rules with exceptions to those rules? i'm sick of seeing "and that rule will work for all words ... except for the following ..."

been messing with language software too. the flashcard vocabulary programs on a Pocket PC are great. especially the ones with audio. not using it yet, but the mobile English-Spanish dictionaries should prove to be excellent as well. have been using interactive software on my Tablet PC too. at least the beginner apps have not required keyboard input, so tapping with a pen works fine. finished one beginner CD and progressed to the next. it is questionable as to whether it is beginner or intermediate. regardless ... i'm lost. finally, i'm judging audio programs as well. some of them are definitely not car friendly. trying to find a really basic one with good production quality and clarity for the car. the others that are of lower production quality or require more concentration will be better suited elsewhere

i've committed to learning Spanish. AI will still remain my primary focus, and Spanish will be secondary. chose Spanish because it is the 3rd (almost 2nd?) most popular language and should be the easiest for me to learn. my assumption is that it will be easy because of its similarity to English, especially the alphabet. also, i will not be at a loss for study material. a quick trip to HalfPriceBooks turned up some interactive software for the desktop, a flash card vocabulary trainer for my Pocket PC, and an English-Spanish dictionary for the PPC. plus the obligatory 'Teach Yourself Spanish for Dummies' book. then i went to Audible.com and got some audio programs to listen to in the car. my 1st goal is to be able to read some Spanish websites. 2nd goal is to watch and understand some Spanish TV. 3rd goal is to actually speak some Spanish. pic below is a poll from Channel 9

we've got a meeting this Tuesday from 5:15 to 9:00pm. we'll be viewing a LiveMeeting presentation of the .NET To Go Mobility Roadshow from Downers Grove, IL. referring to that timetable, you can probably show up a little later than 5:15. the main topic of the presentation will be the Compact Framework. i will try to field questions from the local audience. i think that one of my fellow CF geeks might be there to answer questions as well. of coure all my answers will mention OpenNETCF.org. congrats to Neil, Chris and the CF MVPs for winning a PocketPC Magazine Best Software Award with that effort

1st off ... why in the world did i read a VB.NET book? there is a possiblity i might work on a short VB project, so decided i had better look at it. could not make myself buy a stand-alone VB book, but throwing in game programming made it better. 2nd ... i dont intend to write a game (nor do i play games). reason being is that games have to look good, and i dont have any artistic ability. just bought the book to get an idea of how the stuff works and might possibly use 3D for visualization in other apps.

the book is basically a 2nd edition of another book. it starts out with GDI+, then DirectDraw, and finally Direct3D. it also works in some DirectInput and DirectPlay. most of the chapters involve developing a small game. this is where i think the book shined. it did a really good job of picking what the sample game should be to demonstrate the programming techniques being discussed. the book is relatively short compared to other game books, which i see as a plus. it stays on topic and avoids the fluff that other game books are prone to have. one complaint is that it starts out with some errors in the code listings, but i stopped noticing these after the 1st chapter. another complaint is that it tries to talk a little about AI. from actually trying to teach myself AI right now ... i did not consider what the book offered to really be AI at all. hopefully we will get an AI for .NET book sometime soon? finally, it has a short chapter about porting a GDI+ tetris style game to the Compact Framework

this will expand my thoughts based on current reader feedback. going to kick these (and related) ideas around for a couple months before i commit. more feedback is welcome

1) 2nd discipline. i threw out the possibilities of bio / nano, patent law, and MBA. my thought is an MBA would be easy to obtain. the future paths there would be to a) start a company for outsourcing or b) develop a product. probably would not like myself if i became a conduit for outsourcing. in the long run i do not think it is the right path for the US, not to mention i prefer technical over business. as far as a product, i have yet to come up with an original product idea to pursue. on the other hand, it would actually be relatively easy to just target somebody else's product idea / business model, and then kick their ass technically. patent law isnt really an option, mainly since i despise patents and lawyers. the part that would be good about law is its probably the most difficult to offshore since legal problems are based on a location. the bad news is the US has more than enough lawyers. not to mention it would take a while to attain. nano would also be difficult to attain. the good news is that it is also technical so i would love it. the problem is that it ultimately requires interaction with physical materials which would be hard / expensive to acquire. would have to get my foot in somewhere to seriously do this. the other problem is that this field is based on knowledge as well ... so there is no reason that it wont be outsourced as well to a country with a lower cost of living.
now i did get ripped on (twice) for suggesting an online degree. in one hand i can see how people think online degrees have no merit. on the other hand i've got 1.5 masters degrees ... and they seem to be doing jack towards making employers want to hire me. right now i consider my college education $100K+ not well spent. nor do i hear anything about the degrees obtained from my competitors overseas? so if online degrees have no weight, i'm pretty sure that is equal to the weight that my physical MS degrees have (read zero). not to mention i would learn more teaching myself than actually wasting the time in class. its like if an employer wanted to send me to a week-long technical training class for $2500. i'd be like let me buy a couple hundred in books and just leave me alone for a week, and i'll learn more than anybody that went to the class. thats just the nature of how i learn.

2) 2nd language. all the languages i was considering came up : Arabic, Chinese, Japanese, Russian, and Spanish. i consider Chinese and Spanish the obvious choices. Arabic was a bit of a surprise, but Hindi NOT being suggested was a real surprise? Japanese is in the mix because of Robotics. Russian being suggested was a big surprise? being in the US, Spanish is the choice and is supposedly easy to learn. then i can watch the Spanish channel and figure out why they randomly have hot chicks dancing around, and what the dirty old men dressed in bumblebee outfits are saying. it is also supposed to be easier to learn a 3rd language after picking up a 2nd. internationally, Chinese is the choice based on current # of speakers and projected population.

your comments regarding this are greatly appreciated

(Ray Kurzweil, 1999) had been meaning to read this book for some time. last week skicow posted this link which forced me to read it. Kurzweil is a bad ass. alot of his early inventions were speech and audio related to help the handicapped. then he wrote a book late 80s predicting that a computer would beat a chess champion in 1998. ended up he was wrong and it happened in 1997. so to cash in on that prediction ... he wrote this book :) cracked it open this morning ... and it reads really quick. it is an idea book (not technical). the book is divided into 3 parts : past, present, and future. along that timeline he collects evidence to backup his future predictions for the years : 2009, 2019, 2029 and 2099. most of his predictions are along the lines of machine intelligence outpacing humans, humans becoming more machine like, downloading intelligence to your brain and uploading your brain to a virtual reality. think 'The Matrix' movie ... supposedly this book (or his previous work) influenced those scripts. his ideas scares alot of people and Hawkins mentions (and refutes) this briefly in his book 'on intelligence'. honestly, i think Ray's predictions are technically feasible, but the human element will slow the timeline down; especially considering the evangelical outcome of this weeks election. his new book just came out Fantastic Voyage. it will soon be on my stack of 'books to be read'. it is going to explore the possibility of immortality. the xtians will definitely not go for that. without the fear of death / hell ... how are you going to control the sheep? dare i say ... terror

my sister has Martha Stewart tendencies : arts, crafts, decorating, cooking and such ... not insider trading. checkout the halloween costumes she put together ... thats raw talent

cannot shake this doomsday feeling that the religious right is going to destroy America. so i am going to start preparing myself for the worst. i think the two main things that i need to correct are 1) only knowing computers and 2) only speaking English. i would greatly appreciate feedback on the following. please reply to my comments or email me directly.

1) need to become multidisciplinary. the obvious candidates are medical (nano), legal (possibly patents), and business (MBA). for each of these i would start pursuing online degree offerings. which route would you take?
2) need to become bilingual. pretty confident i can teach myself using avaialable study aids. which language do you think i should learn? obviously i am considering the languages of the big growth centers for the next decade.

all comments will be appreciated. Thanks

hey hey, i'd been searching around for an Artificial Intelligence (AI) portal, since the .NET scene is basically non-existent. the best portal i found is generation5.org. they've got alot of articles, a discussion forum, and updated news periodically. the content ranges from beginner to advanced topics. i HIGHLY recommend this site. generation5 has also (just yesteray) become the home for my own AI articles. the other article in the What's New list should be familiar to early .NET adopters as well, because it discusses Terrarium being used in an Academic setting! how cool is that ...

been getting a number of emails asking for code lately. if you've read bNb PB (pre-blog) then you know i used to give away almost ALL of my code. this year AB (after-blog), i pretty much stopped that. the reason being follows ...

at the end of last year, i interviewed for a large contract that i really wanted. during the interview, i realized they could use some code that i had already written and shared. they were not aware of my code, but got really excited when i told them. anyway, they wanted to pay me a contract rate lower than a full time wage i had right out of school. i could not believe they were serious. about a month later, i started getting a number of questions from somebody using the code. they were real nice, but also very green. eventually i found out that they were working on the contract i had interviewed for. the bad news is this made me bitter and turned me into a code hog. the good news is the other person took months to finish the job i could have done in weeks. not to mention the sample code they sent me was of low quality. the moral of the story is 'you get what ...'

so at the start of this year i made a conscious decision to not give away code for at least a year (as an experiment). and that i would still write articles detailing what steps i took, so that others could recreate it with some effort. sort of 'teach a person how to fish' style. surprisingly, the people that have taken to this approach the quickest are graduate students. there is a handful of postgrad work that is being done using the articles as a starting point, and then taking them further (kind of cool). i certainly try to answer all the questions that they have. back to the issue. at the end of this year, i will reconsider giving away more source code

updated the /aiSomPic application to generate collages of the trained SOM (sample collage below). this makes it easier to visualize how well the SOM has worked. the article contains 2 new collage images representing color and texture of the picture sets used in the article. i've also added some more text in section 4D, plus a collage made from a 5000 images set ... which is just neat to look at at

the professor that helped me with blog aspect of the article was kind enough to link the article. he has a Spanish and an English blogs which i recommend : BloJJ

since my candidate lost ...
now i get to bitch for the next 4 years :)
Bush is welcome to prove me wrong.
which has happened before;
since i was for him last time ...
no comment

(by George W. Smith 1991) this was a Natual Language Processing book. mostly dealing with text, but it also had some interesting bits about speech (both reco and synthesis). started out reading the book very closely, but ended up skimming about half way through. i've attempted to grok NLP a couple of times now, and it is just so dense, that i get bogged down and just cant take it anymore, and have to escape. with other AI specialties, i have to up my math skill set to understand what is going on. with NLP, i have to work on grammar ... yuck! if you read this blog then you should know that my grammar is about at a 5th grade level. sometimes my sentences are even yoda-like due to periodic bouts of dyslexia. i dont even do the grammar check on MS word, because it depresses me. so at this point i'm just training my own brain as i would a neural net. each time i show it the material, it picks up a little more than it did before. eventually it should reach a critical mass so that i might attempt to actually tackle this problem. regardless of my own shortcomings, this was one of the better NLP books that i have looked at

having just read the PalmPilot guys AI book, he talks about beating the handwriting reco problem by reducing the input set with 'graffiti' style strokes. why dont we do the same sort of thing for language? because most of the problems come from all the special cases that can occur. it could have a reduced vocabulary and grammar for computers to understand. somewhere in between command-and-control and dictation. it would be english - ebonics - homonyms - slang - etc... = 'speaketti'

not to brag ... but this is plain cool /aiSomPic. it uses a Self Organizing Map to find similar pictures. it can even search for pictures based off of badly drawn stick-figure representations. the pic below shows it in action. the leftmost PictureBox shows my hand drawn cheerleader picture that i want to search for. the ListView on the right shows the similar images that were returned. the exact image that i was searching for is the 3rd image in that list. the article has alot more info and a short video of its usage

the Tablet has taken some negative hits : here and here. Josh's is losing out to a notebook and a lack of ink applications. a) i think the notebook issue will go away as the price point between the 2 drops. then it will just be an option to upgrade notebooks to have ink as well. b) the lack of applications has 2 parts. b1) MS is mucking up by not making everyone of their apps add Tablet PC support. e.g. i consider WMP 10 not supporting ink a fumble. IE.next having ink support would be big. b2) MS just made it easier for people to develop apps for Tablets by releasing the Recognizer Pack. let me rephrase that ... it is already easy, this just makes it more accessible.

the other article points out that MS is trying to invent unnatural usage scenarios for the Tablet. in rebuttal, i will detail my unnatural usage scenarios. 1) meetings, conferences, user groups : the Tablet couldnt be more natural. it allows me to take notes comfortably and quickly. this is much preferred to typing away at a notebook sitting on my lap. 2) car : for long road trips, the Tablet comes along to keep the passengers entertained. 3) design : its free form nature comes in handy when i am trying to design a new app or brainstorm ideas for the next app. Ink is the ultimate tool for UML. 4) in bed : i actually keep my Tablet near the bed. the slate model is perfect for surfing. it has become my newspaper for catching up on current events. and it has killed TV viewing. 5) study : when trying to learn something new, i usually have to read numerous websites and PDFs. sitting at a desk and reading from a notebook for extended periods of time is unbearable. but i can handle it with a Tablet in a recliner. yes ... the Tablet is making me smarter.

i'll probably have to get a new notebook late next year. at that time, i hope to get a Tablet PC that is in the form factor of a large powerful notebook. one that i can use mostly like a notebook and as a Tablet every once in a while. the kicker is, that i want this Tablet PC to also be able to run as a Media Center Edition. and when my current slate Tablet dies, then i definitely see myself getting another slate too

you ever get one of those bugs that you just cant see? where you know that if you could call somebody else over to your computer to take a look, that they would see it immediately and make you feel really stupid? well ... i've been fighting one of those for the last couple hours. it was actually kind of subtle ... but it was in a core part of my SOM library. and that is where the irony comes into play. SOMs learn through unsupervised training. so basically i had no clue if what it was learning was correct or not. had to dumb down the problem to something that i could see the pattern as well, to know that it was really broken; and not just learning from bad training data. regardless, it is now fixed! disregard my earlier post about the weblog SOM not working ... i need to give that another try now

from previous posts you can tell that i've been playing with SOM neural networks. was able to get the basic stuff to work really well, so spent yesterday trying to tackle something more advanced : Mapping weblog communities. basically it collects posts from a bunch of weblogs and then trains a SOM to visualize the community based on in/out links between blogs. took all yesterday, but i was able to get it to work (sort of). it does create a map of the community, but it is skewed to one corner. i think the reason it is skewed is due to the minimal amount or relationships between the blogs i'm trying to map. also because i did not collect data over a long period of time. the good news is that it does show definite signs of working. for blogs that get mapped away from that corner, then they do seem related to the other blogs mapped nearby. regardless, i'm going to shelve this idea and begin work on a different problem that SOM should be able to solve. this new problem should be a little easier, and the results should be more obvious to see as well

depending on what type of application i'm programming, i am generally concerned with 1 main thing. when programming an app that works with a database, i'm thinking about how fast i can get the work over with and do something more interesting. when programming a mobile app, i'm constantly thinking about performance. web services make me think about if the XML will be interoperable. speech apps make me think about if the Voice UI is intuitive. ... and now AI. in general, when programming AI, about all i can think ... 'is this even going to work'? for the other types of applications, i'm confident i can get them to work, so i can worry about other things; but AI is a different beast. for most of the stuff i've attempted so far ... its basically a leap of faith. i mostly have to write the entire app before i can even begin to see if it might have a chance to work. i'm not entirely comfortable with that. its not like i can isolate the risky parts and tackle those 1st ... its more like the whole thing is one big risk

here is another look at a SOM neural net. this one is used to categorize colors ... which seems to be the 'hello world' of SOM programming. the 1st PictureBox shows the Kohonen net before being run. the 2nd PictureBox shows it after being run and classifying 9 colors. the 3rd PictureBox shows the hot spots of the net where the colors were best classified. the 4th PictureBox shows where the net has the most error because it is transitioning from one classification to another. now that i've got the basics ... i'll try to apply this to something useful

Richard Sprague posted an interesting pic a couple days ago showing How does SR accuracy change over time. if you look real closely you can see the word Sphinx ... which is CMUs open source contribution to speech reco. now we need a similar chart showing the trend to improved accuracy for speech reco on mobile devices :)

decided to go back and revisit neural networks. instead of doing a trained NN with backpropagation ... i'm playing around with the less popular SOM (aka Kohonen). the interesting part is that it is untrained ... and basically learns patterns by itself. the picture below shows a SOM that can learn to draw a grid. the small PictureBox's show its progress as it learns, and the large PictureBox shows its final state

every once in a while, i am in the real world and talking to people about technology. rarely, that person reads my blog. when that does happen, and we end up talking about what i've worked on recently, it becomes difficult to not repeat what i've already posted in the blog. they'll end up saying something to the effect of 'yes, i already read that'. and in the back of their mind, they are saying 'quit wasting my time'.

alternately, if somebody doesnt read the blog and asks me what is going on ... i'll get pissed that i have to repeat myself. to their face, i start reciting the posted blogs with a smile. in the back of my mind, i'm thinking 'dont they know how to read'? and therein lies the true power of podcasting ... promoting illiteracy

still looking for problems to fit some solutions. my solution set currently involves fuzzy logic, hidden markov models, and natural language processing. neural nets are purposefully excluded from the solution set, because i dont want it to become my AI silver bullet.

fuzzy logic 1st. i can see how it is well fit to control systems and rule based systems. but being able to write logic in english-like terms just is not that big of a deal to me. the operators of my language of choice are quite adequate. maybe some 'else if' or 'switch' statements to really spice things up.
hidden markov 2nd. typically used by speech recognition for recognizing phonemes and words. the other standard is to train it on past works and then have it recreate new material based on the statistics of the past work. that is novel, but not very useful. for kicks i applied it to the text to speech program. all it did was randomly pick a starting phoneme, and then chose the following phonemes statistically to form words and sentences of gibberish. something like that might be useful for games like the SIMS, where the characters make speech like sounds.
natural language processing 3rd. i dont know enough to actually code this yet. the first step is to break up a sentence into the parts of speech. this would be difficult for me because i suck at grammar. you should know that if you've read this far ... second step is to determine what the sentence actually means. thats when things get interesting. still trying to formulate ideas from that point on. need to do more research to come up with something meaningful. regardless, this would probably take a significant amount of time to put something together.

so i'm slowly building an AI toolkit and figuring out what technique fits which class of problems. now my new problem is coming up with an application to code applying that technique in a way that solves some meaningful problem. i.e. pairing up AI to application coding is harder than i expected.

my Tablet PC has been crapping out randomly since upgrading to SP2: "Display Driver Stopped Responding" - ialmrnt5. this is happening on my 1st generation Fujitsu slate. happens about once a week. the behavior is it freezes up and stops accepting ink for about 10 seconds, then it displays the dialog below, and if i dismiss that dialog it just keeps on working like nothing happened. i.e. everything still seems to work fine. upon reboot, it does not prompt me to send feedback to MS as the dialog promises. anybody else seeing this? anybody else see this before the SP2 upgrade?

just got a direct email about my WS-Idiots post, in which i attacked people attacking WS-*. i'm going to answer that email publicly ...

regarding my experience with CORBA ... i have programmed with CORBA (check the resume). also did J2EE, but never DCOM. granted it was 5 years ago and i was right out of school (read green). yes, i know its composable. at that time it was very difficult to develop with, and interop was near impossible. it was great at interop within an intranet scenario. over the internet and between partners ... not so good. it probably has gotten alot better since then.
regarding XML-RPC ... i agree that XML-RPC is for simple scenarios and WS-* is for the enterprise. but I think that basic SOAP (without WS-*) is just as simple as XML-RPC. and as your needs change, then you can build on top of that.
regarding WS-* ... i agree that nobody will deploy WS-* without reliability, transactions, security, etc... where did you get the feeling that i did not agree with that? that is exactly why XML-RPC / REST dont have any legs. that is why mobile devices need WS-* for CF, which i constantly bash MS over the head about. i also agree that it will take years for WS-* to really take off.

now my turn for questions ...
do you think the WS-* specs are that complex?
more complex than the CORBA specs?
do you think that CORBA is more interoperable than WS-*?
do you think CORBA can be used in simple scenarios like basic SOAP?
why are the largest software companies pushing WS-* and not CORBA?
the open source argument does not cut it because IBM likes open source and they are pushing WS-*. and there are open source WS-* implementations such as WSE for CF, Plumbwork Orange, WSE for MONO, plus many others outside of the .NET world

i'm supposed to be doing AI and not WS ... but Clemens Vaster just brought up a problem i've come across too : Whenever you start thinking stuff is stable .... had a similar problem when writing the WSE pieces for CF. in my case though it was trying to get it to work with both WS-Routing for WSE1 and WS-Addressing for WSE2. the same thing came up when the WS-Security namespaces changed when moving from WSE2 Tech Preview to WSE2 Release.

he is calling for WSE 2.1 ... but i dont think that is going to cut it. we cant expect everybody to constantly update their WS-* implementations (and apps) every time a new spec comes out. instead the WSE implementations need to be backwards compatible. the SOA building blocks are supposed to be pluggable so you can take what you need and forget the rest. this doesnt just mean which spec ... it also means which version of a particular spec. otherwise ... namespace versioning hell

by Jeff Hawkins (the PalmPilot / Treo guy). this was a book about AI ... sort of. its mostly ideas about how Jeff thinks the neocortex of the brain works. if he is correct and we can recreate that model in soon-to-be hardware, then his thought is that we will be able to build truly intelligent machines. when i 1st started reading about neural networks, i went ahead and read a short book about the brain, thinking that might give me some ideas. but it was a waste of time because the brain book did not say anything about how the brain might actually work, and neural networks only slightly related to what we do know about the brain. the same complaint is made in this book. it also complains about how our current neural networks are way too simplistic to ever become intelligent. also attacks the thought in AI that we should be building something that is intelligent in the same way we are considered intelligent.
Jeffs model for the neocortex primarily consists of a heirarchical structure with memory and predictive capabilities using both feedforward and feedback paths. sort of a tree of auto-associative neural networks. with this setup, it would start out knowing nothing, and it would have to be trained. similar to how we are trained as we are growing up, or how the environment trained our ancestors through evolution. after training it would be able to think in abstract terms to predict its way through the world. this is different than alot of AI concepts that are mostly seeing what is in the real world and then reacting to it. e.g. the robot moving a centimer every hour pseudocode : done processing, move forward, oh crap, something moved, stop, begin processing next movement ... (wait an hour repeat). was hoping that SOM neural networks would be mentioned as well ... but they were not.
so this is the book about the brain, that relates to AI, that i assumed already existed. it was a really good book. the only problem is that it does not come with source code (that was a joke). it is one thing to read about these concepts, its an entirely something else for me to try and program something like this. meaning i wont be writing any code based off ideas from the book. it is targeting bigger problems than i am ready to attempt right now. no, i didnt just like the book because it was a mobility guy writing about AI. actually owned a PalmPilot when i was younger and stupider (intended) ... and could not ditch it fast enough for a PPC. color was just too sexy to resist

by Seul, O'Gorman, Sammon. read this to support the pattern recognition with AI effort. the problem is that you have to get real world data into some form that you can ultimately use an AI technique with. for audio, this has involved working with wav files as well as fourier transforms. for image and video (sequence of images), there are a lot of options : smoothing, sharpening, edge detection, subsampling, line thinning, ... mostly i've just used global pixel and convolution techniques so far. e.g. when attempting the barcode image recognition, i had to transform that from gray scale to a binary format (without loss of data). this book covers many of those techniques as well as line, point, and frequency analysis. it even comes with C code, but i have not looked at it yet; many of the algorithms could be understood through the textual descriptions alone. this book was published 2000, but seems to be much older. regardless, great book, i'll probably be using this as a reference for some time

dont exactly know why i own this book? it must have been an amazon.com impulse purchase? regardless, it was on my stack of books to be read ... so i read it. it has a number of do-it-yourself projects, such as building your own antenna ... which i would never do by myself. also showed how to create a war-driving setup. the problem with that is if i wanted to war-drive, then i would just write the software myself. the only thing that i really got out of it was finding out about power-over-ethernet ... that's cool! at least i got it out of the way, and can move on to the next book

out of the feeds i'm subscribed too ...
thinking-out-loud and The Dev Theologian

Alex Yakhnin posted SOA or where's WSE for mobile devices. you should read it; but the point is that everybody is pushing SOA, but we basically come up empty when it comes to SOA for devices. yes, i wrote the bits that will let you call WSE 2.0 from a device, but you would have to do alot more work to actually host a WSE Web Service on a device. granted you could tie the CF WSE bits together with the Mobile Web Server ... it certainly could be done, but wouldn't be trivial; nor would it perform. i was actually part of a team that did implement a CF Web Service Server (minus WSE) so know exactly what it would take

for my personal opinion ... there does not seem to be a rush. basically, i've only had 1 call for doing professional WSE work in CF (ever). the DIME bits are being used in a handful of places, but nobody is contacting me about WS-Security. even worse, nobody is contacting me about WSE work (period). you might think that writing the WSE bits for CF would make me an expert ... and that i would get contacts for doing WSE on a server? but it has not happened. when i do get contacts about WS work ... these people still think that SOAP, WSDL and UDDI are the advanced standards. they dont even say SOAP or WSDL as one word ... they pronounce each letter individually. from my perspective, WSE has to take off on the server side before it will take off on mobile devices. and in a non-dogfood-like manner, the MapPoint Web Service does not use WSE yet

so what hope is left? i see a couple forces. SOA will not take off until Indigo. VS 2005 is not going to do it because ASP.NET offers very few WS improvements (without Indigo). CFv2 also offers very few WS improvements. Always-connected Mobile devices will push the market too, along with Location Based Services to those devices. Also, the Devices Profile (UPnP 2.0 proposal) will force it to happen. Finally, competition. Michael Yuan posted months ago about Nokia's big plan to beat .Net CF in mobile Web services

by Steven Levy. picked this book up from a post by Softwaremaker after i had just finished Levy's book on Artificial Life. had put together the crypto API for CF to use with WSE, so know how to use crypto, but had no clue of its history. this books filled in those gaps detailing who was behind the different algorithms and how they came about. it also outlines the concerns of the NSA for crypto going public (read terrorism and child porn). it was an entertaining read. also gave me some insight into why my one Lotus Notes friend happens to know a boat load about crypto. it is worth noting that the book was published pre 9/11, so the reader misses out on feelings of guilt. maybe the book will get updated in a couple years to include that and quantum crypto

by Bart Kosko, who is a stud in both fuzzy logic and neural networks. this was an ideas book that mainly related fuzzy logic to politics, science, and digitial culture. it has minimal technical content on both fuzzy and neural network systems, as well as how they can be used together. it does have plenty of thoughts about the future, to make you think. read it because i'm trying to come up with an idea to code using fuzzy logic. can definitely see how it applies to control systems ... but just cant think of a system that i want to control ...

had to kill some time, so saw 3 movies (from worst to best):
sky captain - this was the 1st movie to have its opening sales beat by a video game release (Fable). i can see why. it is a ridiculous title. when i saw the title ... the thought that came to my mind is that we must have run out of available movie titles, and that we would have to start reusing old ones. the only thing special about the movie was that it was filmed entirely with a blue screen. the worst part was that angela jolie wasnt a 'bad guy', and wore alot more than just an eyepatch
team america - my expectations were too high for this movie. yes, it was funny ... but not that funny. i was expecting 'south park movie' or 'something about mary' funny ... this was more along the lines of 'hot shots part deux'. i.e. i laughed more after the movie while remembering certain parts, than during it. and at no part was i like 'im going to hell for laughing at this'. wish that i would have waited for an unedited DVD version. as far as a political comedy ... 'the daily show with jon stewart' is better
friday night lights - i played texas football for 12 years ... and have been disappointed with all high school football movies to date. this one was pretty close to the truth. i read the book when i was just starting high school football, it was basically considered the bible, and a lot of my team read it too. texas football is big business. in my district, football tickets were the only athletic events that were not free to attend. we played at huge stadiums (15K people) with astro turf and almost every game sold out. occasionally we would get helicopter news coverage as well. made the varsity team early, and all of a sudden everybody knew my name. wearing your letter jacket around opened certain doors as well. yes, we really did have signs in our yards too. it really seemed like alot of pressure and at some point i stopped enjoying the game ... did not regain the love for it until playing non-scholarship college ball. anyway ... very well done movie. the only part i didnt like was the Chavez actor was too small. that was the character i emulated ... the student-athlete [aka brains-N-brawn] :)

/ttSpeech is about rolling-your-own Text-To-Speech (TTS) application, showing how to create the simplest speech synthesis program possible. got it to work using managed code exclusively (i.e. SAPI is not used) ... as well as running on a Pocket PC with the compact framework. at best it sounds somewhat like a monotone human, at worst it sounds like 'speak & spell' ... either way, it is recognizable english

(more observations from a coder in a researchers world)
1) where are the AI blogs? they pretty much do not exist. you can take 2 viewpoints here. the 1st is that they dont need them. news does not happen as quickly as it does in the .NET world. in the coder world we have room for endless incremental improvement. the AI world has alot more trial and error. it seems to be much harder to get incremental improvements, and there is alot of backtracking to try something entirely different applied to the same problem. 2nd is that they might need blogs even more. since things dont happen as quickly, its easy to forget about somebody elses work. the 'subscribe and forget' model as i call it would help to keep updated on somebodys work when they come up with the rarer revolutionary idea.
2) what in the world is postscript? actually, i do remember postscript from school ... but PDF has totally taken over. for AI, i had to install a PS viewer for the 1st time in about 3 years.
3) enough with the MATLAB. i've heard of it but never had to use it. just so happens that it comes up everywhere in the AI world.
4) bibliographies. remember how you used to have to do these? they are still all over AI research. haven't these people heard of hyperlinks?

seems odd to me that the internets initial purpose was to serve research papers ... yet many of these researchers dont seem to be using it in a manner much further than its meager beginnings. its basically FrontPage style pages wherever you go. no xml, xslt, flash, javascript, videos, rarely images. maybe they'll retool for the semantic web

where in the world did that come from? was there a demand for it? i'm not positive, but i think another feature that C# lacks and VB.NET gets is the 'My' quick dial thing. would rather get that instead. so if MS plans on forcing a bunch of VB.NET developers to do C#, and MS wants them to bitch less along the way ... then good job; otherwise ... yawn.

while we're on the subject of stuff i could care less about. Team System has a number of tools i'm never going to use. Class Diagram, Unit Testing, Source Control, Code Coverage, Profiling ... good. Modeling services and how they are connected and having it build projects for me with the web references ... WTF? Modeling the logical layout of those services and web apps and the zones they are in ... WTF? i dont want an MCSE using my tool. yes, the single IDE is a good thing ... but dont make it an IDE for everybody. the chances of me using those diagrams ... 0. how about Profiling and Unit Testing on mobile devices instead. or what about Use Cases and Sequence Diagrams. no, forget that ... when am i getting my Tablet-enabled version of Visio which is so intuitive it is like modeling UML on a napkin. the 'UML on a napkin' thing always cracks me up :)

this is significant. the Recognizer Pack can be installed on non-Tablet PCs to provide ink recognition. previously ink recognition could only be tested on an actual Tablet PC device. this should open the flood gates for alot more Tablet developers. Wacom will probably sell a chunk more digitizer boards as well. NOTE that it also installs on Windows Server 2003 ...

the 'HAL 9000' WAV file below is from the homemade text to speech app. a pic of the application is below. i'm pretty happy with how its turning out, considering how very little code is behind it. going to play with it some more and should have an article about it over the weekend.

wav

so i've been working on a custom text-to-speech (TTS) engine. just playing around to get a better idea of how TTS actually works. there are some open source ones out there, but they are more complex than i wanted. i.e. they do alot more processing to try and sound like human voices; whereas i just want something dead simple that might be recognizable as speech, regardless if it sounds like 'speak and spell'. anyways, i 'think' i just got past the hardest part?

... i forget if my blog engine can handle this? if so, there will be a wav file link below that has my computer speaking the letter 'a'
wav

i really hate failure. i am attempting what we will call 'Project X'. this is my 2nd attempt at Project X. wont tell you what Project X is, because if i cant get it to work this time; and then somebody else does, then you'll be like ... "hey, how was Person Y able to do Project X ... but casey couldn't. casey must be a loser." anyway, i really want it to work. been working on it 4 days and still dont have any results to show. had to read for 3 days before i could even start coding. spent all day today trying to figure out the last unknown. getting close, but still did not get it to work. if i can get past this hurdle, then it should only take another day to get the rest of the code put together. even then, Project X could still be a total failure. there is not really any way for me to check to see if is working until it is done. Project X is basically all-or-nothing

so if you dont see a new article within about a week ... then Project X has been postponed (for the 2nd time). at which point you are no longer free to mention Project X. if you bring it up then i will deny any and all knowledge of it. this blog will self destruct. BOOM!

even though my opinion does not count ... you will listen (insert jedi persuasive hand movement here). 1st off, my political party is the anarchist party. meaning i dont have a political party ... because anarchists cannot form groups (see the definition of anarchy). atheists aren't allowed to swarm together for the same reasons. back on topic ... when the debate 1st started, and they shook hands, i think kerry patted bush on the back to check to see if he was wired. bush did much better this time, so i considered it a tie based on how they performed. based on judging them as individuals, kerry definitely wins. bush was jeckyll and hyde regarding international and domestic policies (dis)respectively. i forget if jekyll or hyde was the crazy one ... thus the (dis)? kerry was even throughout the debate. kerry showed his pure brainpower by being able to recall previous names and forming complete sentences with less interruptions. instead of a debate ... i'd like for the candidates to have to take some sort of standardized test. kerry also wins for promoting his website ... you got to like that. bush did mention the internets ... but i doubt that he knows about Internet2? finally, bring back lehrer as a moderator. gibson let bush push him around at least once ... i think lehrer would have been there with the buzzer

this is a recent book put out by 'charles river media'. i've just recently been buying books from this publisher, and mostly like them so far. they seem to get the right author for the job, and they offer topics that you dont see elsewhere. which brings us back to computer forensics. bought this book hoping that it would be mostly technical, so i could recreate my favorite scenes in CSI: Criminal Special Investigations (thats a lie, i dont watch TV; but i think i saw that name on a TV guide?). instead this book was mostly about processes and checklists. it was thorough but also seemed to repeat itself frequently. also very wordy, weighing in around 600 pages ... so i just skimmed it after a while. the main sections that stood out in my mind pertain to data backup and recovery, information warfare (abbreviated IW if you are cool), and analyzing machines for evidence. its probably a good book ... just not the right book for me

let me repeat : i do not work for microsoft. have not been around THAT long, but have been doing .NET for almost 4 years. before that i did Java. when i did Java, nobody ever made the assumption that i worked for Sun. but since i have been working with .NET, i have been downright accused of working for MS an inordinate number of times. is it just me ... or does anybody else get this too?

maybe the people are just mocking me? its their way of asking, why are you evangelizing MS technologies for free? the answer is, i'm really evangelizing myself. any MS promotional material is purely accidental. i just happen to be using .NET because it is the best tool there is right now. i used Java when it was the best thing. when a greater power than .NET comes around, then i will promptly donate my collection of .NET books to the nearest library, and start the next one. if you want a quotable line, '.NET isnt love'

nothing like being an uncultured american. so when foreign people write about me ... i have absolutely no clue what it means. like this (pic below). lets dissect it :

'belove' and a pink'ish background ... GOOD
name in all caps 3 times ... BAD
'Windows Longhorn' in Pascal case ... GOOD
2001 ... BAD
website linked twice ... GOOD
lots of square boxes ... BAD
smiley face ... GOOD

my conclusion is that it must be the work of terrorists and i will continue to live in fear

must say that AI is a different beast than i am used to. i'm used to new technologies such as .NET, Web Services, etc... stuff that has very little info available. such that i could read every Compact Framework book ever written or be the 1st to self-publish an article on some beta product. AI is not like that at all:

1) its been around (off and on) for something like 45 years! i consider the people that have been working on this before i was even born as cheating.
2) in the .NET world i'm used to having tons of web sites with great content to support me: MSDN, codeproject, gotdotnet, etc. there are a number of AI sites, but alot of them have aged. cannot tell you how many dead links i've visited. serious, if i search for something on google, and then branch off from one of those results ... dead links everywhere.
3) and ALOT can happen in 45 years. if you want to find the latest and greatest, you are going to have to filter out a bunch of stuff that used to be the latest and greatest. this also makes it difficult to find the really basic stuff if you are just getting started.
4) the field is HUGE. there are a ton of specializations each having a ton of techniques. then when you start mixing these together, the possibilities seem endless. i feel somewhat lost trying to figure out which AI technique should be applied to which real-world problem.
5) too much research and not enough development. its easy to find a handful of research papers about an AI topic, but if you try to find code ... good luck. if you want OO code ... [crickets chirping].
6) the papers are just loaded with math. its cool that they want to show off their skills; but if they want developers to actually USE their techniques, instead of just having other researches read about it, then how about some pseudo code and some diagrams. typically they do not provide their code that goes along with the paper.
7) there is no Microsoft. if i want to bitch about .NET ... i have a target. there is no such target in the AI world. AI is everywhere ... and nowhere. about all i can bitch about is my brain not coming with a reference manual. stupid brain :)

in some recent article i mentioned 'hidden markov models' ... and admitted that the name was familiar, but i had forgotten the concept. so i spent most of today brushing up. the 'hello world' of HMM is to feed it a body of text from a specific author and then use that to generate a new sentence that has a likely probability of occuring. ported some C code over and ran it against the transcript of the presidential debates. it was not as funny as i had hoped for. if i told it to start with the word 'osama', then it did generate 'osama bin laden'. could also get Bush's corpus to say 'it's hard work' by making it start with the word 'it's'. HMMs can be applied to speech recognition, natural language processing, music modeling, and spell checking. now i'm trying to come up with a different project idea in which it would make sense to use an HMM. NOTE the use of the word 'corpus' above ... i've been reading too many academic books about AI ...

this news.com headline from yesterday got me riled up: Where's the simplicity in Web services?. basically its about people bitching that there are too many WS-* specs and/or they are too complex.

1st) let me politely say that these people are idiots, and they should not be running your business.
2nd) these are probably the same people that said SOAP was too simple to begin with 4 years ago; because it lacked specs for large files, security, reliability, etc...
3rd) the other option for language and platform independence is CORBA. if you think CORBA is simpler than WS-*, then you deserve a boot to the head.
4th) the other option is REST. if you dont think you can make very simple SOAP requests without doing WS-Security and the advanced specs, then you deserve a kick in the nuts.
5th) how complex is XML? all the WS-* specs are about is getting the XML to match a schema. making XML look the way you want it to is about as simple as it gets.
6th) MS and IBM are making toolkits to even hide that little bit of complexity from you. we've got .NET, WSE, and Indigo tomorrow. can you say Don Box?
7th) i've stopped working on the WS-* implementations for CF in my spare time because i dont consider it a challenge anymore. you look at the WS-* spec and then make your XML look like the spec. if you get lucky, then you'll have to write some actual code with logic. if you get even lucker, you might have to deal with an interop issue ... its grunt work

a simple article about doing video gesture recognition : /aiGesture

if you dont believe that most music on iPods (or any mp3 player) is stolen ... then you dont have a clue. if i got an iPod, it would weight about 10 pounds more after i was done putting stolen music on it. do the math; how long have mp3 players been out - how long have online music stores been available. that said ... DRM will never work. RIAA needs to come up with something else. the funny thing is ... i've gotten so sick of the music industry ... i rarely listen to music anymore. yes my friends ... coding in silence

todays book was: Implementing Biometric Security (Chirillo and Blaul). it was a quick read because it was not what i expected. seeing the word 'Implementing' in the title made me think that it would give me some low level details on how to write my own biometric algorithms from scratch. instead, this book is about 3rd party tools and how you can use them in your biometric setups. so basically i just read the conceptual stuff and skipped everything that had to do with a specific implementation. it does go through a number of physiological and behavioral biometrics explaining the pros and cons of each. it also explains what concerns people might have about each type of biometric.

with some image processing techniques i picked up from my AI studies ... got the /barCode image reader to work on a Pocket PC using an HP Photosmart Mobile Camera. its using the Bistia camera wrapper and Compact Framework v1. all you do is take a picture of a barcode, and then it does the recognition routine and displays the result in a MessageBox.

the scenario is that you would use this with a Smartphone (or PPC Phone Edition) at a bookstore. then it would send the ISBN to the Amazon.com Web Service to lookup what their price was. the pic on the left shows the image captured from the camera with minimal preprocessing. the pic on the right shows the transformed image used for recognition, and the MessageBox displaying the ISBN #.

just finished 2 more, and both are highly recommended: AI Application Programming (M. Tim Jones) and AI Techniques for Game Programming (Mat Buckland)

Application Programming is a tour of various 'weak AI' techniques. i know 'weak AI' sounds negative ... but it is not. the book covers: Simulated Annealing, ART, Ant Algorithms, Backprop Neural Nets, Genetic Algorithms, Artificial Life, Expert Systems, Fuzzy Logic, Hidden Markov Models, and Intelligent Agents. there is a chapter on each one of these techniques. each chapter does a great job of introducing the concept and giving a simple example to start off. the minimum amount of math is provided, and if necessary, examples of the math is given. then the chapters move into a more complex example with C source code to explain what is going on. each chapter ends explaining what problems that technique is best applied to and a list to other resources. e.g. the GA chapter started out with an example on Genetic Algorithms but ended up with an example on Genetic Programming. this is a wonderful book for programmers that want to start learning AI

Game Programming goes more in depth with Genetic Algorithms, Neural Networks, and combinations of both. it is mostly about AI, and the examples just happen to be about Games. it starts out with basic Windows Programming in C++ (just skip those chapters). the 2nd part is about Genetic Algorithms and how you can evolve your programs to learn algorithms. the 3rd part is about Neural Networks. it starts out with MultiLayer Adaline Networks at 1st, and uses genetic algorithms to train them in a somewhat unsupervised way. then it uses BackPropogation Neural Nets and supervised training. finally, it introduces EANN (Evolutionary Artificial Neural Networks). EANNs use evolutionary techniques to determine what the topology of the Neural Net should be. e.g. how many layers, how many nodes in each layer, what the connections should be, and their weights. this was new to me ... and honestly i'm not ready for it yet ... but i do know there is already a C# implementation called SharpNEAT

1) thumbnails. icerocket wasn't the 1st, but i like how they dispaly thumbnails of webpages. it helps for weeding out crap but it is even better at helping me remember which link takes me where
2) site info. both icerocket and A9 have site info from alexa.com. it is not crucial for search but it sure is interesting to look at
3) history. how A9 keeps my past searches, because i frequently search for the same things over and over again. i really used to love how copernic did this, but just got sick of having to re-install the fat client when i was traveling so much
4) multi search (need a better word?). really like how A9 will return image results at the same time as web results
5) fat fingers. how google will fix misspelled words. this is particulary useful on Tablet PCs because my handwriting is so piss poor. now i just write like crap and google fixes it
6) local. the google local service kicks ass
7) clusters. how the new clusty engine groups results into relevant topics. some of my clusties are

Blog
Speech
OpenNETCF
Tablet PC
Smartphone
Web Services

hell ... that is exactly how i've been positioning myself! in 3 to 6 months, hopefully AI will be a clusty as well

now i want a single search site to give me all of this. actually, give me the option for all of this search functionality, and then let me personalize it to how i see fit

that is exactly what i consider this to be: H-1B visa limit for 2005 already reached. there are still many unemployed US programmers (myself included), but companies cannot seem to snag up H-1Bs quick enough. but they have advanced degress ... er, um; i have 2. a majority of the H-1Bs that i've met on my nomadic travels were primarily here to learn the market and then head overseas to run outsourcing operations. this is odd, because i've never been contacted for a telecommuting job by a US company. my own rates would certainly be lower to telecommute ... but it has not come up once. another thought, why do we have limits on importing jobs and not on exporting? it seems like if you were to have limits on importing or exporting, then you would limit exporting. thus, you could import as many as you wanted, and that would keep more of the money within the US. not an economost ... but that seems like common sense to me?

bill gates does not know it, but he has been my mentor. you know how your parents would give you advice: get good grades, go to college, get a job. well, after my parents limits were reached, i turned to the most obvious adviser ... the richest man in the world. my choice of going into computer science in the 1st place was partly because of bill. more recent choices have been based on his nudgings as well: .NET and Web Services, Security, Seamless Computing (mobile, web services, location, speech). and here is bill pushing CompSci and Artificial Intelligence. that will give MS (and Google) a handful of AI researches. but go figure ... i'm teaching myself AI. so i've got all those checkboxes covered ... but the job market is still crap. if you are unemployed, it is a zero-sum game; and its not much better if you are employed and making substantially less than what you used to. even though i despise typing this, i would certainly not advise US students to study CompSci. i'm under the impression that if you are a US programmer, and really want to program, then you need to leave the US and move to where there is a lower cost of living ... then you can compete on fair terms. because the competition right now seems to be solely based on cost ... and not skill. finally, i'm looking for a new unsuspecting mentor

now that i'm older and wiser ... words of wisdom: people live too long. i'm already starting to fall apart. my eyes are going and my hands crack periodically. once i cannot code productively anymore, you might as well just take me out back and put me down. that would have the benefit of making more room in the environment for something else to evolve that was better suited. there are some genetic algorithms that use this sort of logic in their code. they implant every member with a 'cyanide pill' that goes off after so long to ensure that the algorithm can keep optimizing.

i'm under the impression that the medical industry keeps us alive as long as possible, to acquire more money from us. especially when considering costs of pharmaceutical drugs and medical insurance. the older you get, the conditions you have just get worse and cost more. instead of cancer and diabetes, i'm for reverting back to heart attacks. quick and unexpectedly ... that is the way to go.

reflecting like this ... because i'm nearing the dreaded 30. which brings me back to living too long. now that we live so long, a 'mid-life crisis' does not cut it anymore. we should really be allowed to have a couple of them. i'm talking 'one-third' and 'two-thirds life crisis' respectively.

[ink recognition] XXIX

isf

i used to play around with MS Agents. one misuse of the technology was to make them read scripts. 4 years ago, i had them read one of the Bush/Gore debates ... so i went ahead and put up the transcript of the Bush/Kerry debate from last night: /agents. BTW Kerry kicked Bush's ass.

(ASIDE) bad news: i had a blue screen this morning. good news: i could not remember the last time i had a blue screen. i'm talking months if not close to a year. and i'm not exactly easy on my machine :) wasted CPU cycles ... what's that?

voracious reader here. e.g. my Amazon.com bill was just over $3000 for books last year. there is probably another $1000 on top of that for books bought in-store (i need to get a sponsor). so i'm perfectly fine shelling out $40-$50 for a programming book. but the problem is, alot of AI books (on Amazon) are in the $75-$125 range! ... and that is putting me through sticker shock. i really do not see myself shelling out that kind of dough for a book, especially with some of the top rated books being around 10 years old. luckily, the older and more popular books can be found on ebay and such for about half price. also been raiding the collections at Half Price Books. been able to find a number of the classics by visiting the local stores. these books seem to have a low turnover rate ... so you can pick them up cheap. been meaning to, but i have not scouted out the local off-campus used book stores as well. also been buying the 'AI for Game Programming' books from Amazon. they seem to have the right mix of theory, mathematics, and application that a programmer can easily understand. you just have to read it and replace the terminology with application developer terms. e.g. 'troop path algorithm' is 'business rule optimization' and 'non player character' is 'intelligent agent'

had a bunch of down time today so put away a couple books. 1) Made-up Minds (Drescher) is another thesis turned into a book. this one is about building a general purpose AI to try and learn in the manner of the stages of Piagetian child development. i liked the concept and it had some good ideas but its not something i would attempt anytime soon. 2) The Shape of Actions (Collins and Kusch) is about classifying the human actions that machines can and cannot do. it was heavy on describing actions and light on any sort of AI topics.

this article (/aiTabletOCR) is about my 1st foray into the world of Artificial Intelligence programming. it uses a Neural Network to do stroke-based handwriting recognition of characters on a Tablet PC. half of the article is about Neural Nets, and the other half is about applying one to this particular problem. it has some basic NN code that i ported over to C# from C, and links to more robust libraries with source. there is also a video clip of it being used.

anyways, this is my new kick. most new articles should be along the same lines as this one for some time out ...

yesterday, i attempted to do OCR from my handwriting on a Tablet. that did not work so well. the results were far from satisfactory. so i stopped going down that path and started down the stroke-based path. the difference being OCR works with images and pixels, and strokes work the x/y coordinates of the pen. beefed up my test app until i could find what looked like a recurring pattern. if i mapped the angle that the pen was traveling to a wave form, then that seemed pretty good. so i shaped that data until the neural net could work with it ... and the pic below shows the very 1st test. 4 characters with no errors ... on the 1st test! ... i was floored :) will have to see what its limits are and then see if i can tweak it to make it better

another tricky thing about neural networks (and AI in general) is how you get the data into some format that will be useful. typically, the input data needs to be an array of values between -1 and 1. so how do you get that array of values? well ... it varies for each thing that you are trying to do:

for barcodes, it is easy, because everything is already black (1) or white (0). if you have an image that is of lower quality, then that gives you the grey values between 0 and 1. transforming it into a sine wave helped too. for the EAN format, the array length is 95. granted ... i have not been able to successfully read low quality barcode images ... yet ...

optical character recognition is simple too. you just divide up an image into areas of pixels, and then its value is that areas grayscale. next, you just normalize the grayscale between 1 and -1. the array length is then how many areas you divide the image up into e.g. 10x10.

speech recognition was entirely different. the WAV format gives you an array of values which you can immediately plot on a graph. in that format though it was basically meaningless. if you do an FFT transform though, then you can divide it up into segments and produce the array of values by summing (similar to OCR above). then those values just have to be normalized between -1 and 1.

now i'm attempting handwriting recognition (like the Tablet PC). first, i tried OCR with minimal success. now i'm trying to work with the actual stroke data. plotting out the strokes movement as a wave looks promising. also tried FFT, but it looks less promising, and fails altogether on some inputs. there are also some OCR-esque techniques that take the strokes into consideration that i might try later.

anyways, this was not obvious to me before starting AI ... but it takes ALOT of massage work to get the input data into some format so that you can even apply the AI technique. so now i'm constantly having to look into image processing techniques and mathematical transforms. normalizing between 1 and -1 is trivial, but the thing i still dont understand is what to do with variable length inputs. e.g. if i'm trying to recognize a short sound vs a long sound. their input arrays will be of different lengths ... and i dont know how to deal with that ... or if the NN is able to handle that automagically for you?

went ahead and hooked up the neural network OCR to run on my Tablet PC. it had been using images of characters rendered with GDI+ for training and test cases. now it is using my own handwriting for both. have to enter some letters and numbers, then train the neural net. then i can write on the lower textArea and have the neural net do the recognition (instead of the Tablet Recognizer). results aren't great, but it does have minimal success (pictured below)

aka lessons learned. mucked around with the neural networks again today, by adding some error feedback mechanisms. basically, this lets me judge the total error in the system as it is being run, but it is most helpful in training.

there seems to be multiple levels of training that you can attain: first, is to just train it until it properly matches all the inputs to the correct output. second, is to achieve the same goal, but to feed the inputs in a random order. this makes sure that it trains evenly across the inputs. third, is to achieve the same goal, but to keep running it to reduce the error. this seemed to help it handle input with noise better. granted, there is a point at which you can overtrain and make it too specialized, thus making it worse at handling noise.

what else ... training sucks. you waste a bunch of time waiting for it to train. highly recommend using an NN that can de/serialize itself once it has been trained; although i've been tweaking it a bunch of different ways and retraining to see if a certain setting works better. that part also sucks ... tweaking the NN. some examples are: how you format the input. the range for the random weights that you initialize it. what transfer function you use for activation. if you use bias. what learning rate and momentum you use. if you use a different learning rate for the hidden nodes for the output nodes. how long you train it. what minimum error is acceptable. that is a lot of different things you can muck up. i have not done it, but am considering making a 'training harness' that i could use to let it try out all these different permutations, and keep score of which setup did best. then i could just let it run overnight until it came up with the best fit. sort of procedural genetic programming if you will

when they search the job sites and spam everybody that fits their keyword search. i.e.

1) when they dont check the timeline. e.g. they ping me for 'java' ... which i have not done for 4 years.
2) when they dont check context. e.g. i get contacts for java 'webMethods', when my resume has the word 'WebMethods' as in ASP.NET.
3) when they dont keep a list of whom they have already contacted. e.g. i get an email saying 'hello, found your resume on monster.com', and then 15 minutes later 'hello, found your resume on computerjobs.com'.

this just reinforces in my mind that coding is a commodity. there is really no reason for me to keep making myself a better developer, because of the disconnect that exists between me and potentional customers ... that disconnect being recruiters. you've got a high tech company on one end, then you've got technical workers on the other ... and smack in the middle, somebody that does not know technology. usually 2 of them. 1 that faces the client and one that faces the consultant. that is like me asking my mom to go search for information on neural networks that is going to help me out. chalk another one up for the stupid category

[propagate is a funny word] today was spent hacking another neural network. this time i tried the one from pseudonym67: NN.NET The BackPropagation Network. 1st off, pseudonym67 has a number of great AI articles on codeproject ... recommended. on the off chance he/she reads this, i want their email addy. the NN code he provides has 3 different types of NNs: Adaline is too simple, so ignore it. BackPropagation is the one just about everybody uses, and i used today. SOM (or Kohonen) which is just plain weird because it uses unsupervised training, so look at it for cool factor. the thing that sucks about the BackProp NN articles is that they use lame examples with only 1 output and a max of 20 inputs. basically all these can tell you is a boolean response. but the code provided is better than that and can do more complex scenarios. had to muck with the network code a little, but was able to get it to do the same OCR with the network i was using yesterday.

the code was different, but the 2 BPNs were inherently the same. when running, the network from yesterday had the additional usage of Threshold. this one did not have that, but it could optionally use Bias. when learning, yesterdays used 2 different learning rates for the output and hidden layers out of the box. this one can support that as well, it just takes a miniscule effort. this one also adds the concept of Momentum. yesterdays had some error code which was not entirely hooked up. todays has more error code ... but i dont entirely understand what it is supposed to be telling me yet

Kurt (who is a bad ass BTW) just posted Patent Absurdity. he has specific examples of patents that MS has applied for in New Zealand. in one aspect ... i agree. software patents are mostly BS (i mean that with the utmost respect). but i dont really consider MS the bad guy on this one. 1) IBM is the bad guy. they have more freakin patents than anybody. there are online patent databases you can search which are flooded by IBM. i actually had a cool idea once, and every related patented was already owned by guess who ... 'International Business Machines'. 2) IBM has been doing this for a long time now, and MS has just recently gone on a patent frenzy (by rewarding its employees with black cubes they can stack into pyramid forms). if i recall correctly from a past article, MS wants 3K patents this year (up from 2K last year); but i think IBM already had OVER 3.5K patents LAST year. 3) MS has to get these patents to protect their own ass. e.g. EOLAS. regardless, i consider it to be a modern day nuclear arm race, and at some point we'll be like 'these things could really blow everything the f#%* up', and then we'll come to our senses and start dismantling them (unstacking the black cubes) ... as well as trying to keep them out of the hands of terrorists. e.g. SCO.

it looks like my own neurons are starting to form in the brain to allow me to understand artificial neural networks. just got a basic OCR type app to work! its using a Back Propagated Network with images of characters as input. currently the exact same concept of this article: Creating OCR Apps using NN. the problem with that article is that it does not provide the code to the neural network. so i stole that idea and recreated it using this code: C# NN Class. not too hard, but the reason i'm jazzed is that i understood why certain things were not working and how to change the NN code appropriately. might consider tricking this up and turning it into an article

out of the push to learn AI ... neural networks (NN) are the most interesting to me right now. there are a couple .NET implementations available so i dont really plan on writing one from scratch. on the other hand, i did port some basic NN C code over to C# to play with. within the field of NN, there are many different types. today i experimented to see that a Perceptron can learn OR but cannot do XOR. believe this was the 'cause' that 'effect'ively beat the connectivist AI approach down when it was getting out of the starting gate. also saw that a trained Adaline network does not handle noise in the input well. these are the building blocks to the BPNs which seems to be the dominant flavor of NN. also considering Hopfield and Self-Organizing networks as well

to turn things around. here is Kirk telling us not to post politics on our blogs. actually, i find it more offensive for somebody to tell me what not to do. i find political posts easier to ignore.

so since he is asking for it ... my political views are ... undecided. i thought the DNC was lame, and thought the RNC was even more lame. it's just the same BS from both groups ... and there is so much freakin lying (and spin) going on that it just makes me sick. i dont like that Kerry cant seem to stick with his decisions, and i dont like that Bush is all about fear. instead of picking who is going to do the best job, i feel like it is coming down to which one do i dislike the least. aside ... i think Mark Cuban would make a kick ass president

Jason Haley got a pic of the infamous Live Bait Machine. awesome! i had never seen one of these things before, but this one was along the way on my drive to/from work. did a triple-take when i 1st noticed it, and then started asking around. my boss came up with the best theory, that it is for the hardcore fisherman to get their bait before the store opens.

i was dying to know what was inside too. my theory was that it was a perfect little biosphere. all it had inside was water, dirt, some plants, and creatures of the bait derivative. of course sunlight was needed, like the self contained ecospheres. it had sensors and such to create the proper amount of bait. if the bait was not purchased, then the creatures would die and become part of the food chain. then nobody would ever have to come and restock the machine ... it would just continue on indefinitely. the only problem would be if the bait evolved ...

you've been HAACKED posted a link to Ishkur's Guide To Electronic Music v2.5. now i cant stop playing with it.

started out with rap music, but then had to stop when i actually listened to the words; because i did not promote any of that. so i opted for music that mostly lacked words. been listening to electronic music for 15 years now. yes, when starting out, it all sounded like the same song. it took me a little bit before i could recognize when one song started and another began. granted i will never be able to differentiate the different types of electronica to this level of granularity of genres.

its funny that the site even shows my main progression: Jungle Tab - Rave - OldSkool - Jungle - Drum N Bass. also branch down from OldSkool to DarkCore - HardStep - TechStep - Darkstep. did not even know the last set were genres, had always just considered them more agressive Drum N Bass.

the sad part is that my musical tastes have not changed for the last 5 years, so i am still listening to mostly music from the late 90s. you know how your parents listened to oldies stations ... i'm already there

started reading about image processing yesterday, in relation to pattern recognition. today i revisited the /barCode article to try and apply some of those techniques. the code is now a little more elegant, but the main benefit is it more robust. the problem is that i still dont have enough 'soft computing' techniques in the code to handle lower quality images. the real problem is that i have not internalized the techniques yet. i'm building up a skill set of AI techniques, image processing, etc... but i have not solved enough problems with these tools to know when to apply what technique where. e.g. would fuzzy logic or a neural network work better in this situation? e.g. what type of neural network would make sense here? a specific example is that one of the image filters i applied to the barcode just made things worse. probably going to have to backtrack to some simpler problems before moving on. the good news is that i'm finding this extremely challenging ... and having a blast!

just finished 2 books: Genetic Engineering: Dream or Nightmare (1999) and Nano: The Emerging Science of Nanotechnology (1995).

the gen. eng. book was mostly a doomsayer book. read it because i'm just beginning to think about writing genetic software programs, but i wanted some real-world genetic knowledge. surprisingly i agreed with a majority of the points that were being made about how gen. eng. could muck up the real-world. the main theme seemed to be that we could figure out the genes, but that does not necessarily mean that they will behave in the wild. it was just a little too much 'mother earth is an organism' for my taste. overall, the field is interesting, and there is an off chance i might pick up a newer book on this topic with a more balanced opinion.

the nano book was more interesting, albeit older. it was mostly about the people behind nanotechnology and the critical discoveries that drove it, with some futuristic ideas for good measure. picked this book up because i ultimately want to write code to control atoms, or code that controls the nanomachines (this want is probably further out than my gen. sofware desire). surprisingly, with the earlier publish date, this book did mention gen. eng. a little as well, as a field that might help make nano a reality. it repeatedly mentioned the assembler; a microwave type device that we will eventually use to form atoms into whatever we want. but my favorite concept was the nano-fog, out of which items would seem to materialize out of thin air. definitely will pick up a newer nano book, especially after this news.

both technologies have related 'red button' 'destroy the earth 20 times over' type scenarios. for gen. eng. it was creating a 'super virus'. basically a virus that could not be stopped. for nano it was 'gray goo'. the concept was you create a nano machine that creates 2 identical copies of itself, which create 2 identical copies of themselves, repeat. so that ultimately all matter would be turned into these self-replicating machines. scary stuff. the good news is that we have not blown ourselves up with nuclear bombs.

i got some good answers (from MS) regarding the IP of the Dictionary. basically MS requests that i do not post the WordList because of a license that they have, and they are looking into this. so as a courtesy, i will continue to not post the WordList. it does seem to be OK that i derived the WordList to come up with the words that did not have lower-case representations, to better improve my Tablet experience. now i feel better about the whole situation. BTW i probably will write some more Tablet articles, cant help it ... coding is a sickness :)

not a game player ... but for some strange reason i fired up the Internet Games that come with XP. Checkers, Reversi ... and such. the human behavior that really caught my attention is that i was never able to finish a game ... unless i lost. if i was winning or there was a momentum shift in my favor, then my opponents would just quit the game. very few times was i able to make it all the way through the game to fully win, instead of winning by the opponent resigning. but if i was losing, intentionally of course :), then my opponent would typically play the game to the end!

the results above were at the default 'beginner' level. if i went up to the 'intermediate' level, the opponents were better sports and i could actually win some games. the funny thing is that if they lost, then they would always go find a new opponent; but if they won, then some of them would stick around to try and beat me again. i did not try the 'expert' level because i did not want to lose :(

this is just one reason (i've got others). here is an article about Secure Windows Mobile Development and Deployment that recently came out. NOTE that it does not mention the System.Security.Cryptography wrapper for CF, or the WSE 2.0 WS-Security bits. you've got to be kidding me. MSDN only linking to MSDN ... stupid

the job market still sucks (if not getting worse), so i've had alot of time to read. here is what i've kicked back lately:

Expert Systems and Fuzzy Systems - think it was a thesis turned into a book. it did not really go into detail about Expert Systems or Fuzzy Systems, but talked about how to use them in a complementary manner. wasn't too much of a waste of time.
GIS: A Visual Approach - reads like a long powerpoint presentation (split text and graphics). quick read to keep refresh the jargon and concepts in my mind.
Bots, Spiders, and Intelligent Agents in C++ - bought this years ago and am just now getting around to read it. reinforced the concepts that i had taught myself by building my own spider in C#, but did not introduce anything i didnt figure out on my own. should have read it when i bought it, then it would have saved me some time.
Catalog of Tomorrow - great near-term technologies book. covered many different topics with the proper amound of depth. got kind of gloom and doom towards the end about the environment. would read one of these again in a couple years.
Knowledge Web - mistankingly thought this was going to be about knowledge webs. instead it was a history of technology that tied a bunch of seemingly unrelated events together. figured out this was a waste of time about 1/3 through and stopped.
Artificial Worlds - intro book about Artificial Life. dont remember anything bad about this book, but some of the ones below are better.
After Thought - another book on Artificial Life. this one was good, but I think I liked another one below even better.
Thinking Between the Lines - another thesis turned into a book. about feeding a computer a bunch of dictionary terms and definitions, and then getting it to be able to answer questions from that knowledge base. went too much into linguistics for my tastes right now. might have to pick this one up again later because it definitely has some worthwhile info.
Implementing Security for .NET (70-340) - this was for MCSD .NET. probably the best exam study guide i read. i recommend it even if you are not going to take the exam.
Artificial Life - my favorite book on Artificial Life. this one had a great progression from experiment to experiment. will probably have to skim this one again to reinforce the major ideas of A-Life, as well as to spark some of my own ideas.
Introducing Genetics - short book with alot of worthless graphics. useful for a quick overview of genetics.
Introducing Mind and Brain - short book with alot of worthless graphics. decent overview of the brains. let me find out that i probably dont need to read anymore books about the brain to help me with AI.
Programming DirectShow - was the 1st book about programming DirectShow. mostly C++ but had one chapter for .NET. i'll probably end up doing some DirectShow stuff in the future and will revisit this book.
Construcing Intelligent Agents using Java - 1st part of the book is about the concepts, and the 2nd part is specific applications. wished it would have been more about AI and less about agents.

Nano - just started a book about nanotechnology. seems like the author is very positive about the technology, so i will like the book
Genetic Engineering - just started this book too (yes, i typically read a couple books at the same time). this author is very negative about the technology ... so i probably will not like it. it is good to know both sides of the story though

i'm trying to come up with a list of file extensions that render TEXT/HTML on the web. right now i've got: htm,html,xhtml,shtml,php,php3,phtml,xml,asp,aspx,jsp,cfm,cfc,cfml

what else am i missing?

Dare posts: The RSS Sky is Falling...Again. just so happens that i have a solution to the bandwidth problem. all people have to do is voice their opinions to become controversial. start blogging about religion, politics, and sex (in that order) ... that will get people to unsubscribe and save you bandwidth.

actually, i have no clue how many people read my blog (because i have not written that functionality, nor am i compelled to). even when i read somebody else's blog that shows they read mine ... i'm hardly ever listed in their blogroll. possibly too controversial? here is a funny stat though ... the number of times that i've used the word 'sex' in my blog per month:

Jan - 0
Feb - 12
Mar - 11
Apr - 6
May - 5
Jun - 4
Jul - 1
Aug - 0
Sep - 3 (as of now, including this post)

NOTE the gradual decline has not been intentional

saw that Incremental Blogger is waiting on the /tabletDic article. well ... i'm still waiting to hear back from MS about whether i can post it ... hopefully monday. i'm actually kind of kicking myself for pulling it. it would be fun to get some legal mumbo jumbo from MS. of course i would have posted it too :) anyways, i'm not exactly motivated to write any more tablet pc dev articles ... go figure

Mark Harrison points to some MS products with fingerprint readers. that is cool for authentication purposes ... but another problem to be solved is how to de-authenticate. i.e. prove that i am not somebody. this would solve the sort of problems where we have authors doing their own anonymous book reviews on amazon. e.g. the error message would be, you cannot review this book because you authored it. its too easy to get around it now. either be anonymous; or if authentication is required, create a 2nd alias to authenticate as. the figerprint biometric does not solve this either because i could just use a different finger. maybe a DNA keyboard would do it? its only a matter of time until DNA will be signed by a certificate authority ...

'gmail invites' got your attention ... didnt it? too bad, because i'm not giving them away. they are going to sit and rot. first, i jumped on an email invite from a friend. now i feel somewhat obligated to be nice to that person because they were nice enough to give me a gmail invite. well, i dont want to give away invites and trouble you with that same burden. the scenario is:

inviter: i'm so cool
casey: no you are not
inviter: HEY! you ingrate, i got you gmail
casey: doh! you win. you are cool

the reversal is if i give you an invite, then i expect you to praise me endlessly. second, i dont want to be googles drug dealer. i expect that people push invites on their friends and families first. gmail is definitely questionable when it comes to privacy issues; so i'm not going to be signing up everybody i know. e.g. i would sell gmail to complete strangers, but not my sister. this also involves having to choose favorites ...

sister: thanks for the gmail invite
casey: you are welcome
mom: how come i did not get a gmail invite
casey: i ran out
mom: you ingrate. blah 9 months blah ...

third, i've got other things to do than deal with complete strangers. they've got the #1 search engine on the web, why do they need me to spend my time doing their advertising. dole them out periodically from their home page and help even out the digital divide. also, the people that actually use google will be getting gmail invites, instead of me sending them to people that search with msn / yahoo / icerocket. admittedly, i dont know any of those people ... that was just theory.

finally, if you saw the title of this message and blindly sent me an email begging for an invite, then i will post your email address on my blog for the email spiders to harvest and pummel you with ads for sex and drugs.

a friend tipped me off to certification Transcript Sharing, to prove that you really are certified. as opposed to hanging the paper certificate up in your cube like a diploma :) i like it. are universities doing this too? i think they should. then you could just link this stuff directly from your resume. also, i would like an option where somebody could browse to it without having to enter an Access Code. just let them pass the Transcript ID as a query string to directly link it.

Transcript ID 658671
Access Code 12345678
(has not been updated with yesterdays test)

70-315 (5/08) 874
70-316 (5/14) 830
70-320 (5/22) 968
70-300 (6/19) N/A
70-340 (9/07) 828

so now i'm an MCSD .NET. i've never been certified before ... and i used to make fun of people with certifications (so you can make fun of me now). got certified mainly because recruiters are idiots. they know what MCP, MCAD and MSCD are ... but they have no clue what an MVP is. ultimately it is just more flair for the resume ... which cost about $750 to get (including books and exam fees). more costly was the month of time that was spent studying. my study material consisted of the MsPress study guides (for all tests) and the free questions at CertYourself.com (for many of the tests). people swear by the Transcenders but i could not justify spending the additional money. that said, Transcender does have some free sample tests that i wished i would have found out about sooner. i've seen that a couple people have got dual certifications in both C# and VB.NET. yes, i think recruiters are stupid enough to not know It's The Runtime Stupid ... but i'm not going to drop the cash to take 4 VB.NET tests that are inherently the same as the C# tests i've already taken; and would be trivial to pass having passed the other. also, if a company does not realize that i could do VB.NET at the drop off a hat ... then i dont want to work for that company either. granted, i would bet that some recruiters would pass me up because i will not have MSCD in VB.NET on my resume. did i say stupid?

so that was my motiviation for getting certified ... will let you know if it pays off in the future. the immediate gain is minimal ... if not a loss. by studying for the tests, i learned a few esoteric things, but nothing that i will probably ever use in my day to day work. it did make me learn some things that i would never look at on my own, so it is good to have a better understanding of those concepts, although i will promptly clear most of that info from my short term memory. if i have not already used it in my almost 4 years of .NET ... then i'm probably not going to use it in the next 4. basically all i've done to this point is 'prove that i know what i already knew' (part of the reason i write developer articles too, although i hope those articles prove i know more than what certification proves). the only test that i really thought was valuable was the 70-340 Security test, which i also thought was the most difficult. so right now, i learned a little, spent $750, and spent a month of time on it. i wonder if that month of time would have been better spent teaching myself something new, and writing (a) developer article(s) in the process? i'm certain the TS would have preferred that i spent that cash on her instead :)

what would i change? 1st off, if a person passes the test, then show them the correct answers for the questions they missed! the person spent the money (and the time), if they proved they know it well enough to pass, then train them to be better by showing them what they did not get right. it will only improve the developer ... which should be the ultimate goal. the only logic i can imagine for not showing the correct answers is to hinder the people with photographic memories that memorize the tests for people that provide exam study guides. all you are doing is slightly slowing these people down when you could be doing developers a greater service. 2nd, let people specialize. there is too much in .NET for a person to know it all: Win Forms, Web Forms, Services. make the specific tests harder, instead of doing easier tests across the board (i.e. jack of all trades, master of nothing). also offer more specialized tests, like one for ADO.NET and another for the Compact Framework. this would give people more choices for the core requirements. 3rd, overhaul the electives. the Security elective is great. the Biztalk exam needs to be updated to 2004, kill off Sql Server for Yukon, and kill off Commerce Server entirely. 4th, get rid of the trinkets. i dont what any certificates to hang in a cube or pins to put on my notebook bag. that stuff stopped being cool when i turned 16. 5th, spend some of the money that you use to promote certifications to promote the MVP program instead. i consider MVP more valuable than MCSD, but not enough people know about it

test was only 30 questions in 1.5 hours ... i was expecting more questions. about 10 more questions would have been right. my score was an 828 (700 to pass). that might be my lowest score out of all the tests. appropriately enough, i think this was the hardest exam out of them all, because there are so many interdependencies of security. e.g. CAS doesnt even take place if the app is running with Full Trust. what is the authentication type in both IIS and the web.config. the user can just bypass your app and go directly to the file system. etc. the questions on the other tests were specific, and you could just concentate on that one part, ignoring everything else ... while this one you had to take into account many more factors at once. luckily, this test had one of the best study guides. read the book in 2 days, took the practice tests the next day, and the actual test the day after. i learned more studying for this one than any of the others and recommend this book even if you are not going to certify. knew there was alot of security built into .NET, but when you get it all at once, it is almost overwhelming. this book even had a couple lessons that were not explicitly on the exam objectives ... and at least one of those topics was on my exam. the CD practice test was better than previous MsPress offerings too. it had more non-multiple choice questions, as well as 300 total questions to practice with. i think the previous MsPress sample tests only had 100 questions. overall ... this was my best exam experience. even though i think my score sucked, i think i got the most out of this exam. the other tests were basically proving what i already knew ... so i didnt get much out of them. that said, i'm still going to run as Administrator, following the practice of what i call 'computing darwinism'. meaning if i'm stupid enough to destroy my computer ... then i should not be operating one

err, umm ... i sort of got in trouble. was asked (by an MS representative) to pull the /tabletDic article today. since i'm friends with that person ... i did. the problem was that the article gave away a majority of the words in the Tablet PC Dictionary. this lexicon is protected by MS. the key points being 'MS legal' and 'reverse engineering EULA' ... just like Paycheck with Ben Affleck; minus me getting paid, having my memory erased, and Uma Thurman :(

supposedly somebody will review the article and determine if / what i can re-post from the article. should know something later this week

anybody else ever get in trouble for something they have coded? similar things like this have happened a handful of times to me now. i've gotten cease-and-desist letters, hate phone calls, hate emails, DMCA warnings, and shut down by ISPs. my domain name is still blocked by some proxy servers. no death threats ... yet :)

somebody wake up Scoble ... i said 'Tablet' :) just announcing my 3rd article in a row concerning Tablet PCs. the new one is /tabletDic. it details my efforts to determine the words contained in the Tablet System Dictionary ... because MS would not tell us. this dictionary is used for ink recognition as well as alternate reco results. the end product is a 146K word list of unique words the dictionary contains. the article also explains why the Tablet PC is not great if you use all lower case letters, where RSTNLE comes from, the longest words, what dirty words are there, and some important words that are missing ... such as 'blog'. the dictionary is somewhat proprietary info ... so we'll see if i get in any trouble ...

why doesn't it have Tablet PC gestures built in? these guys saw the need. also in one of my older articles. by now ... every app coming out of Redmond should have features specifically for the Tablet PC. i.e. the next version of IE should include gestures for browsing

going to explore for a couple days. i'll post an article about the Tablet PC Dictionary when i get back. as a precursor i'll say that the default dictionary does not have any of George Carlin's 7 dirty words :( how does MS expect me to properly express myself without cussing?

Julie Lerman points out that the DoesYourAppThinkInInk contest is wrapping up.

i just spent the last couple days putting an entry together and made it into the /tabletSign article. it is a signature biometric for logging into web sites with a Tablet PC. an interesting part is that the user does not have to use a text-based password ... they could draw a hieroglyphic ... or whatever. it also uses WSE security to communicate back to the web site, so that the biometric info cannot be tampered with or viewed (i.e. Xml-Signature and Xml-Encryption). it does all this without requiring the user to change their default web browser security settings

disconnected (again) around 5 PM yesterday. EarthLink (ISP) sent TimeWarner (cable provider) another disconnect request yesterday ... for no reason. i can call EL and tell them to reactivate it, and they do immediately ... but it takes up to 72 hrs for EL and TW to synch. how is it that my internet providers dont have the ability to synch instantly? are they using FedEx or what? this really does not give me much faith in them. should be a secure web service call (while i'm on the phone with them) and then everybody is happy. this is what happened this time ...

casey: [calls EarthLink, wait time is 30 minutes]
casey: [hangs up and calls TW] did EL disconnect me again?
TW: yes
casey : [calls EL, waits 30 minutes] why did you disconnect me?
EL: whoops! our fault. its fixed. that will take 72 hours to resynch
casey: you call that fixed? call TW to get it connected now
EL: ok. i'll call right now [casey hangs up]
[next morning ... still not connected. calls EL]
casey: call TW and get me connected now
EL rep: i cant
casey: let me talk to your manager
EL man: i'm the manager
casey: call TW and get it connected now
EL man: let me put you on hold [15 minutes later]
EL man: ok, i called them and i'll have to re-submit this
casey: is it going to be connected immediately?
EL man: it could take up to 72 hours
casey: then what the hell were you doing when you called TW?
EL man: i'll submit this now, then you can call TW
casey: [waits a little and calls TW] EL called, please connect me
TW: please disconnect your modem and turn off your computer
casey: WTF! did you even listen to me?
TW: we have to go through this 1st
casey: [goes through network for monkeys process ... does not fix it]
TW: please hold one second [transfers call to EL]
EL: welcome to EarthLink
casey: WTF! [hangs up and calls TW back] EL called, please connect me
TW: this is messed up, they sent another disconnect command today
casey: WTF!
TW: let me call them now to get this fixed [on hold 20 minutes]
TW: just spoke to EL. they are sending the reconnect now.
it will take an hour. i will call you back in an hour.
casey: Thank you [2 hours pass. casey calls TW]
TW: we have to call EL again [30 minutes on hold]
[casey has to leave, TS takes over]
[hour later, TS gets the internet fixed again]
(total disconnect time - 23 hours)

amazon: usually ships within 24 hours. now i dont have an excuse for finishing MCSD. the Biztalk 2000 elective needs to get updated to 2004 next, and then SQL Server 2000 to Yukon. the Commerce Server 2000 elective needs to go away ... never to return.

longest word (34 chars.): SUPERCALIFRAGILISTICEXPIALIDOCIOUS
2nd longest (31 chars.): DICHLORODIPHENYLTRICHLOROETHANE
3rd longest (30 chars.): HIPPOPOTOMONSTROSESQUIPEDALIAN

this is just a guess: i know explicitly that you cannot pInvoke from an embedded WinForms control without requiring additional permissions. my assumption was that calls to System.Security.Cryptography.*CryptoServiceProviders would pInvoke the native CryptoAPI ... and then throw a SecurityException. but that is not the case. both SHA1CryptoServiceProvider and TripleDESCryptoServiceProvider are working ... without additonal security. i think it is because they are not really doing a pInvoke; instead they are doing [MethodImpl(MethodImplOptions.InternalCall)]. so hashing and symmetric crypto seem to be fine. makes me wonder if pub/pri key encryption will work? the additional burden that might make it break is that the RSA key pairs are stored in the registry

a little more info. the main changes i had to make to get my WSE bits to run in an untrusted environment involved my choice of crypto. i chose a path where i would be able to switch all cryptography over to fully managed classes (the rest is just shaping XML to the WS-Security specs). the best choice for this was to encrypt and sign with a UsernameToken. the crypto for this path involves Random number generation, P_SHA1 for key generation, SHA1 and HMACSHA1 for signing, and TripleDES or AES128 for encryption. what i expected is that everywhere there was *CryptoServiceProvider, i would have to switch it to *Managed, because *CryptoServiceProvider ultimately calls the native CryptoApi. i changed some of them, but missed a couple ... and it still worked, but i dont know how? my expectation was that those calls would throw a SecurityException. will have to so some more digging

from my rant below ... i thought that i had coded something cool ... but needed to test it on the internet ... well, it has been tested ... and it works!

just browsed to a web page which contained an embedded WinForms control (like an applet). that WinForms control has a button on it. when clicking that button, it makes a signed and encrypted WS-Security UsernameToken SoapRequest to a WSE 2.0 WebService, which returns a signed and encrypted SoapResponse. the real kicker is that it made the call without needing any additional security. i'm talking default Medium-trust Internet-zone baby! this works for embedded WinForm controls, as well at No Touch Deployment (NTD) apps. maybe ClickOnce, once that happens. did it by refactoring the WSE code i wrote for CF. i'm not holding my breath until we see a java applet do WS-Security

9 AM [casey coding at his computer]
10 AM [internet connection goes down]
2 PM [finishes reading a book, calls Time Warner]
casey: internet broken. fix it.
TW rep: its EarthLinks fault.
[calls EarthLink]
casey: TW says you broke internet. fix it.
EL rep: whoops, let me change that. all better.
[casey works on Tablet PC dictionary attack] 4 PM [still no internet. calls TW 2nd time]
casey: EL says they fixed it.
TW rep: no they didnt.
[calls EL 2nd time]
casey: TW says you did not fix it.
EL rep: no, we really did fix it. wait longer.
[casey codes something really cool. needs to test on internet] 7 PM [calls TW 3rd time]
casey: EL says they fixed it the 1st time.
TW rep: it might be fixed, but they have to call us.
otherwise 48 hours til re-connect.
[calls EL 3rd time]
casey: TW rep says you have to call them to fix it.
EL rep: i cant call TW. lets redo what didnt work last time.
could take 72 hours.
casey: TALK TO MANAGER NOW
EL man: hello i'm the manager
casey: YOU FUCKED UP. 72 HOURS UNACCEPTABLE. FIX IT
EL man: lets call TW [makes 3-way call to TW]
TW rep: your names not on the account. the TS has to call.
we cant fix it until she calls.
casey: [SILENCE! DEEP BREATH ... 10,9,8,7,6,5,4,3,2,1 ...]
this is to protect her from terrorists trying to get her connection fixed?
TW rep: [silence.]
EL man: [silence.]
8 PM [casey calls TS and TS calls TW 4th time]
TS: i'm the TS. fix it.
TW rep: we dont have the info we need.
we cant fix it until tomorrow around noon.
[TS calls back TW for the 5th time ... they fix it]
thank you TS

[ink recognition] 5 new Power Toys click here

isf

been doing an attack to discover the words in the Tablet PC dictionary. my current estimate is that it contains >150K unique words. alphabetically sorted, the 1st word is $US ... which i think is appropriate. to varying degrees, i've brute forced words from 1 to 6 characters in length. started a brute force for 7 character words today which is going to take (oddly enough) 7 days to complete. using a different method to discover words with longer characters. expect to see the results in about a week

been catching up on past .NET Show episodes. the Lonhorn Fundamentals episode mentions "the Devices Profile for Web Services specification detailing how to connect smart devices via Web services and preview developer kits"

why am i just now hearing about a preview dev kit? had hoped that implementing WSE2 for CF would put me on some radar screens ... but it hasn't. i REALLY REALLY want that dev kit. somebody hook a brother up

Stephen Ness, Mobile Ink Jots 3: Ink on the Web

keep stumbling across articles about people scared to install SP2. i'm certain these are the same people that complained about XP not being secure. right about now i want a super evil virus alert to come out. i'm talking of the 'del c:\*.*' variety. then maybe these people will realize what they should really be scared about. boot to the head

a couple days ago, Chris blew off the /tabletWeb article. to a point ... i agree. it will definitely take a significant amout of time for ink on the web to be adopted, and IE needs some updates to better support it: embedded WinForms controls need to load faster, be easier to debug, be able to interact with the page (each other?), and be able to render alternate HTML text while controls are loading and if the controls do not load at all, etc. even though it seems that i've had a meager # of people interact with the ink-enabled pages (maybe it will increase after SP2 gets pushed out more?), i'm sort of compelled to write another ink-in-IE article. but i need a little more ... so when is the next DoesYourCodeThinkInInk contest? nothing like a possible bribe to push me over the edge :)

so the real problem is i have way too many ideas to code them all. want to do MLS, but cant get it to install. a significantly different ink-in-IE article might be cool, but i just did one and is still TBD if people actually ink on it. too soon to do CF or WS again. maybe something entirely different ...

borrowing a word i picked up from Sam Gentile ... FRICTION! i've been trying to get MapPoint Location Server installed off and on for the 2 weeks. this has got to be the most frustrating install i've attempted in years. the Tablet/Speech article i just put out was mainly just a ploy to take my mind off of this. it is a constant beating of starting an install, getting an error dialog, looking at the docs and knowledge base articles, fiddling with your server, and then attempting the install again ... to either get the same error dialog, or a new one. dont get me wrong, there are installation docs; they just have some holes, or point to KB articles that are only relevant to a degree. from the developers perspective ... i just want to get this thing working so i can develop against it. i do not want to setup Active Directory, DNS, Groups, Users, Certificate Services, SSL certificates. could not care less about the security on my LAN. if you do require all that, give me an installation walk-through. something so point and click stupid that a monkey like myself can do it. then i could spend my time playing with the API. there is a newsgroup for it, and i can see that other people have successfully installed it ... i hate those people :) there are also alot of installation questions :(

i get a number of recruiters that contact me ... but they dont really want to find me work, they just want me to help them find somebody else. the last message was something like "dear casey, do you know somebody that has your experience in web forms, web services, tablet pc, security, compact framework ... but is less senior than you?". that is a polite way of them asking for somebody that knows as much as me, but will work for less money so they can take a bigger cut. they'll find somebody else that does not know as much, but is cheaper, and will not even submit my resume. needless to say ... i dont respond to these types of messages.

Julia Lerman posted Web Doodling

MCE (Media Center Edition) - dont have any decent hardware to load the OS on. dont watch enough TV to warrant buying some either.
PMC (Personal Media Center) - they have the CF CLR on them, we just cannot access it. dont want one until i can code it.
SPOT - i want to be able to write programs to run on the watch themselves. the OS is basically a mini CLR and a HAL. that very concept gets me excited. also want to write programs to send data to the watch. dont think there is an API available to hobbyists in either situation? dont want one until i can code it.
Telematics - been meaning to put a computer in my car for some time now.

this is the latest article /tabletWeb. it is about using the Tablet PC Ink API, along with the Speech SDK, to create web pages that you can interact with using a pen and / or your voice. got a video of it in use that shows it working much better than i expected it would. also, some other examples of doing ink on the web. the TS's favorite is a control that lets you graffiti over my picture (she turned me into sherlock holmes)

anybody know of a web cam that is decent at making videos off a monitor? preferrably something that i could pick up at Best Buy / Circuit City. i just tried the Logitech Orbit, and it could not focus in on the text being displayed. need this to create some demo vids for the latest article. 'camtasia' will not work ... nor will it do it justice

started playing around with some MapPoint stuff ... but got stuck :( just so happens that somebody else is putting together an article that should help me get unstuck ... so i'm waiting on that. in the meantime i started playing with some Tablet PC stuff. got it working better than expected today ... going to trick it out tomorrow, and get it in article form within a week

after installing SP2 on my XP Pro machine, it was taking really long to initially connect to a web site. this was particularly noticable over 802.11g, but i was also getting lag when wired. checked my event log and saw that i had hit the 10 connection limit a couple times ... but installing that patch did not help much. but i did just go through 'add / remove programs' and wiped out all the existing 802.11 drivers and internet connection helpers that came pre-installed with my notebook from the factory ... and that seems to have fixed it ... i hope

trying to make the site a little more user friendly. should have reduced the need to do a horizontal scroll on the blog page. the comments page text box should render a larger size in Firefox. the mobile page now has a blog tab and should auto-redirect based on device as well. plus some other minor changes.

the comments page now supports Ink-in-IE. if you have a Tablet PC with SP2 (and .NET 1.1) then you will be able to write your comments in ink. if you are not on a Tablet, then it will render HTML and images.

[ink recognition] please duck creek dint overflew

isf

had asked about SqlCeResultSet working on Tablet PCs. Bill came through with the answer that it will be ... along with some great posts on working with it.

now the question is ... how does Bill get his questions answered by MS? i've got a bad streak of asking questions to MS, and having them say they'll get back to me ... and then never getting an answer :(

GDI+ in Microsoft .NET

isf

isf

the final version is available for download! also, i discoverd that SP2 updates the OS to Tablet PC Edition 2005 (not 2004, as i posted previously)

i've installed on 6 boxes without a glitch: 2 XP Homes, 3 Tablets, and 1 XP Pro. used the same 400+ meg download from the MSDN subscription for each

my only complaint is on the XP Pro box. i'd been using it over wireless-G to a Microsoft MN-700 access point; except it has gotten noticably slower. just updated the MN-700 firmware ... no improvement. reset the access point ... still slow. going to have to wire this box and see if it improves. also need to unwire some of the other boxes and see how they do.

finally, anybody know what \Windows\System32\DrvrSrvcDll.exe is? SP2 is asking if i want to unblock it ... and i dont know ...

now that SP2 is out the door, Tablet PCs become more respectable. SP2 significantly increases the usability of Tablet PCs, by upgrading them to Tablet PC 2004. my chicken scratch print made the original flavor of the Tablet PC basically unusable for me. it was way too frustrating and part of the reason why i haven't been developing with it recently (used to be a Tablet MVP). i have been running an alpha of Tablet 2004 though, and it is MUCH MUCH better. so if you have a Tablet now, get SP2 ASAP. and if you tried out a Tablet in the past, and passed, then you should give them another try.

now what we need is the next version of the Tablet SDK to come out. it has been in the works, so i'm hoping it is released shortly. once that happens ... expect to see a number of new ink-enabled blogs

just announcing the latest articles:

/cfWSE2 shows how to call almost all of the WSE 2.0 samples from the Compact Framework (sample code included). you will need to look at the code, since the programming model does not match WSE

/cfAES is an enhanced crypto library for the Compact Framework. it includes all of the algorithms for the full .NET Framework v1.1 ... plus the current Whidbey Beta 1. AES is particularly useful for /cfWSE2 above. all you have to do is include the library, add a '#if AES' directive, and you can use this library to call WSE 2.0 using the default symmetric alrgorithm of AES (instead of having to change it to TripleDES)

/speechMulti is about Speech Server 2004 for doing speech recognition and text-to-speech on your Pocket PC with the Pocket IE Speech Add-in. plus a walkthough using the Speech SDK. also goes over how you can wrap it with the Compact Framework to provide SR and TTS within your CF applications

/cfBCL are the results of comparing the assemblies of Compact Framework v1 to CF v2 in Whidbey Beta 1. quickly shows the new namespaces, types, and methods. also breaking changes, and what is unique to CF (not on the desktop)

[off topic] the video page has alot of new links too

the TreeView should be working for FireFox now. i'm actually writing this post from FireFox. just to let you know ... i dont like it; so this is probably the extent of my testing. please let me know if you have any other problems

my blog has been mostly broken for the last couple months. just so happens i'm in TX through the weekend (where my server serves) and am getting to fix some things. permalinks should be working now, the comments page should load faster, and the archives are back. FireFox support is slightly better. it used to be just a black screen. that is fixed ... but the TreeView is still broken. there are some other things i will try to do as well. anyways ... if you have any complaints for other things to fix ... now is the time to shoot me an email, and i'll see what i can do

Sam brought it up. then Bill dropped some code. ... it is cool to have something that the full framework does not [evil smile]. but is that really the case? SqlMobile is supposed to work on Tablet PCs too. does that mean that the SqlCeResultSet will ultimately work on a Tablet PC (running the full framework) as well? or will the devices use the same underlying database, but require a different code base?

these are the types that exist in CF v2 for devices, but not on the full framework for Whidbey on the desktop (Beta 1)

Namespace - Microsoft.WindowsCE.Forms
Namespace - System.Data.SqlServerCe
Type - Microsoft.VisualBasic.DefaultInstanceProperty
Type - Microsoft.VisualBasic.MyGroupAttribute
Type - System.Net.IrDAEndPoint
Type - System.Net.Sockets.IrDACharacterSet
Type - System.Net.Sockets.IrDAClient
Type - System.Net.Sockets.IrDADeviceInfo
Type - System.Net.Sockets.IrDAHints
Type - System.Net.Sockets.IrDAListener
Type - System.Threading.ThreadTerminateException
Type - System.Windows.Forms.AutoScaleMode
Type - System.Xml.XmlQualifiedNameTable

these items are in CF v1, but not CF v2 (Whidbey Beta 1):

Type - System.Text.MLangCodePageEncoding
Type - System.Data.PropertyAttributes
Type - System.Web.Services.Protocols.SoapDocumentServiceAttribute
Method - System.Data.SqlServerCe.SqlCeInfoMessageEventArgs.ErrorCode

if you want me to develop with your product ... then do NOT do the following

these are the new namespaces in Whidbey (Beta 1) CF V2:

System.Data.ProviderBase
System.Data.Sql
System.Drawing.Text
System.IO.Ports
System.Messaging
System.Runtime.InteropServices.ComTypes
System.Runtime.Remoting
System.Security.Cryptography
System.Security.Permissions
System.Xml.XPath
MS.Internal.SchemaValidator
MS.Internal.Xml

some comparison of Compact Framework v1 to v2 (with Whidbey Beta 1 bits):

size on disk:
v1: 2 megs v2: 3.6
public types:
v1: 800 v2: 1,200
public methods:
v1: 13,500 v2: 22,400

from Mark Cuban's blog: Watch Out Google….here we come!

so google has made search profitable, putting a big target on their back, and now everybody wants a piece. bunch of copycats (and inbreeding) in this industry. (dont take that wrong, Mr. Cuban's a bad ass). guess everybody is still scared to do something new. or else new ideas are hard ... and coding is easy

SR is Speech Recognition. this involves the same trick explained a couple days ago for doing TTS for CF. just embed a web browser into a CF app. this one is a little different (than TTS), because you have to pass which grammar to use and kick off the recognition automatically. then you can speak and the web browser will end up with the text that was spoken. the only problem is that the Whidbey WebBrowser control for CF will not let you retrieve the HTML that is returned. so you cannot get the answer from the WebBrowser and back into the main application. that needs to be fixed. an article for this will be posted late next week.

i have this internal struggle involving where i should direct my time. i'm always torn between reading a book about some type of programming i have no clue about OR do i read more about software engineering. e.g. should i read about XQuery, AI, Programming DirectShow OR should i read about TDD, Design Patterns, Agile Programming. that said ... i choose XQuery, AI, DirectShow 99% of the time (wider is better). i've never been on a project and had a client say, 'i am so glad you know Design Patterns, UML, SCRUM, flavor of the day' ... but they have come by with questions about XSLT, Crypto, RegEx, WS-* ...

friend #1 sent me this link (Great Hackers), and thinks my next employer should read it. actually, i've worked a total of 8 weeks this year ... too busy coding to look for a day job. if the TS wasn't spending all my money, then i would have already called it a year and submitted my 04 taxes

A - B

it is out. i just ran the unit tests for the WSE 2.0 Compact Framework bits against SP1 and everything passed. yeah

been upskill'ing (stealing Chris's word) with Speech Server 2004 this week, and putting together an article. should finish it this weekend. that will make 3 dev articles to put out late next week! the 2 other articles are 1) how to call WSE 2.0 from CF and 2) enhanced crypto for CF

i'm subscribed to ~200 blogs now.
these are the ones with unread posts...

Scobleizer - 399
MSDN Just Published - 558
Microsoft community blog posts - 846
Weblogs @ DotNetJunkies.com - 3325
Weblogs @ ASP.NET - 17211

the bolded folder names in Outlook make me feel stressed

this is a little hack i dreamed up at MobileDevCon, and i just now got around to building. it involves doing text-to-speech (TTS) in a compact framework (CF) app using Speech Server. all i did is create a speech-enabled web page that will speak whatever you pass as a query string. then i created a compact framework app with a hidden WebBrowser control. whenever you want the app to speak, all you do is pass the query string'd url to the web browser. it loads the page, then gets speech server to do the TTS, and then the web browser plays the prompt using the PIE speech add-in. it would take a little work, but i think this could also be modified to use Speech Server to do speech recognition for CF. kind of cool

BOOM.wmv (5877 KB)

... and sometimes 'y'. if you enter 'kc', it (Speech Server) will read each letter individually, i.e. 'kay.see'. as soon as you add a vowel, it tries to read it as a word. e.g. 'kac' will be read 'kack' and not 'kay.aye.see'. so i entered the alphabet with consonants only, and this is what worked: bcdfghjklmnpqrstvwxz. note that 'y' is missing; because it would read it as a single word otherwise ... which sounds like a bunch of clicks and whistles from the !Kung

Speech Server will read 999999999999999999 as "nine hundred ninety nine quadrillion, nine hundred ninety nine trillion, nine hundred ninety nine billion, nine hundred ninety nine million, nine hundred ninety nine thousand, nine hundred ninety nine". add 1, and it reads 1000000000000000000 as "one zero zero zero zero zero zero zero zero zero zero zero zero zero zero zero zero zero zero". i discovered this testing to see if it would say sextillion

i liked it (but have never been an Asimov guy). Frank Herbert was my science fiction (Dune) Messiah. [warning : spoilers] i liked the fist biometric. also the part when the girl was trying to get the CD player to do speech reco, and was saying different words, because she did not know the grammar (been there, done that). the TS pointed out that somebody would definitely hack a robot to make it commit crime. the TS also pointed out that working older-model robots would not be stored away in train box cars. and the black guy being prejudiced ... classic

alright ... Speech Server is kicking me square in the balls right now. how in the world do i get my deployed Speech SDK app to work with Speech Server? i'm going through the docs and not finding them particularly helpful for the Pocket PC PIE Speech add-in scenario. i'm using the SES only, and not TAS ... so this should be pretty simple :( i've posted my current questions to the newsgroup: microsoft.public.netspeechsdk ... and now i'm holding my breath (because i'm basically stuck)

also, when i run the SpeechMaps sample from the desktop IE with Speech add-in, why do i get prompted to trust the Speech Server. the desktop is powerful enough to do all Speech Reco and TTS locally ... why is it going to the Speech Server at all? shouldn't it only be going to the speech app web server?

this thread on Don Box's wiki is worth noting. Scoble linked it. my thought is that the device guys are on their own for a while to do WS-* support (on any platform). once we see more web services (on servers) start using the WS-* specs, then the need will trickle down to devices. and now try to name off some public web services using WS-* specs. i can think of one: the Microsoft.com web service, which provides no real value in its current state, and only uses 1 small piece of WS-Security (WSE 1.0). i dont NEED that web service for my device. now the MapPoint web service is compelling for devices ... but WSE isnt being used. it looks like the Pocket IE Speech add-in does web services to Speech Sever 2004 as well ... but no WSE. so this is me calling Microsoft out that they are not adequately 'dog fooding' WS-*. and bring back HailStorm ... couldn't resist :) [segue] it is interesting that out of my WS-* implementations for the Compact Framework, the only one that really gets used is WS-Attachments. this is funny since DIME will be deprecated for MTOM. i've only been asked once (over the last year) about the WS-Security stack ... which does do XML-DSIG and XML-ENC. [segue] actually ... maybe some competition will speed up adoption too: Nokia's big plan to beat .Net CF in mobile Web services

where do baby dolls come from?

i've been mostly training my web spider by giving it negative reinforcement. e.g. if it gets a movie i dont like, then i set a flag that this was bad. this technique has been working ... but it is slow. just a couple days ago i took a different approach, and i added positive reinforcement, to steer it towards good content. now the spider gets so lost in good content, that it hardly picks up any bad stuff at all. this has been working so well, that i even took the 'training wheels' off. meaning i had been purposefully limiting the scope of the spider ... so as to not take up all my time training it. i would basically only let it get 100 movies each time, of which about 25% would be bad. now it has been getting about 1500 daily, and only about 1% are bad. i'm wondering if i even need to train it anymore?

here is a gem from the Speech Server sample readme: "To reduce recognition latency, after installing Speech Add-in for Microsoft Pocket Internet Explorer, establish an ActiveSync connection with the Pocket PC device and delete the file labeled mscc.mtfp. This file is located in the Windows directory on the Pocket PC device."

WTF? i hate magic comments!... google (nothing) groups.google (nothing) searching the Speech Server CHM file (nothing). looking at the mscc.mtfp file (binary). so why in the world does deleting the mscc.mtfp file reduce reco latency? and why is it installed in the 1st place?

been fighting with Speech Server 2004 the last couple days. it had been winning ... but i finally bludgeoned it into submission tonight. yesterday was an install fest: Windows Server 2003, IIS 6, Windows Updates, Enterprise Instrumentation, Speech Server, Speech SDK. i just want to use it from a PPC with the PIE Speech add-in (over WiFi), so i do not have a dialogic board. Speech Server comes with 1 sample that uses the MapPoint Web Service. it sucks as a sample because it has too many dependencies. e.g. i want to get the sample working now, but i have to wait 3 or so days to get my MapPoint user and pass. also, it has config info for working behind proxy servers, which is fine if you are behind a proxy server; but i'm not, and i cant find an intuitive way to get it to ignore those settings. would have preferred a simpler app (in addition) just to do some speech reco and TTS

Sam Gentile points to Matt Powell pointing to a public pronouncement for 26 WS-specifications being royalty free! i was pissed off about this just a couple weeks ago when i saw that i needed to sign a license to implement the WS-Security spec. now i am happier :) kind of interesting that MTOM is on the list, but DIME and WS-Attachments is not. WS-Addressing (WSE2) is there and WS-Routing (WSE1) is not. also, where is WS-MetadataTransfer (to be used by WS-Discovery)? i 'think' WS-MetadataExchange is different? and at the very bottom: 'To inquire about obtaining a license to any of the specifications listed above send an e-mail to stdsreq@microsoft.com'. i've implemented approximately 8 of those specs ... maybe i'll email them and let you know what happens

coonasses.wmv (10743 KB)

this made me think of chris for some reason ...
beaver.wmv (24472 KB)

beatbox.wmv (6237 KB)

word up to Jef, aka thinking-out-loud. i got to work with Jef early in my career (when i was incompetent). he probably does not even know it, but he gave me advice on my very 1st side project (skincycle.com) to use ASP instead of JSP. that was the flex point that led me away from java web development, and ultimately helped me to jump into ASP.NET. next, when i use 'err, umm' in text ... i entirely stole that from him. also, he was the 1st person i met that did not watch TV, which i adopted too. finally, he is the 1st developer i met BB (before blogs) to start blogging AB (after blogs). Thanks Jef.
still waiting for another Jef(rem), who mentored me early on, to start blogging ...

cool ... my WSE 2.0 stuff for OpenNETCF 1.2 made the TheServerSide.net. i really need to get the associated article online (about how to use it), err, umm ... but i broke my server and cannot upload files. whoops :) i wont be able to fix it for a couple weeks.
doing a little more cleanup on the extended crypto lib for CF. i stopped relying on the CryptoAPI for MACTripleDES, which seems to have a bug on CE, and just did the MAC myself. ended up being really simple, so that is working now. still need to take another cut at PasswordDeriveBytes.
found a series of articles on fuzzy logic by pseudonym67 on CodeProject.com (he also has some on neural networks). i've been looking at those, as well as reading some other AI books.

i am ALMOST compelled to get a new iPod. what i consider the killer feature is the ability to listen to audio books at different rates of speed (particularly faster). that combined with iPod's integration with audible.com might just put me over the edge (and i could care less about Apple). can Portable Media Centers do this?

ProtectedMemory for v2 pInvokes into SystemFunction040 and SystemFunction041 ... whatever those are? and i doubt those functions exist on CE :) all ProtectedMemory does is encrypt blocks of memory you pass in, so i cooked up my own internals using DES. this should be OK since ProtectedMemory is not supposed to be used outside of a single machine, although it can be used cross process. e.g. you could encrypt the memory in one process, then pass that encrypted memory to another process ... which would then decrypt that memory space. i THINK i have that working (with a hack) ... but i did not test it.
and ProtectedMemory allows for a SecureString. all i did was port the code from Hernan de Lahitte. it seems to be working, but i need to do a little more testing to be sure.

implemented ProtectedData real quick. all it does is provide static Methods to Protect() and Unprotect(). you can pass in entropy, which i will have to extend my CryptProtectData pInvoke to support. not exactly sure how to support ProtectedMemory yet ... have no clue what pInvoke the full framework is doing? or if it is available on CE? not sure how useful it is either, because it requires the memory to be in 16 byte blocks.
Rfc2898DeriveBytes was a bit tricker, but it is now working too. RSA reccommends doing 1000 iterations to generate a key ... this is a BAD idea on a PPC. it took way too long. will have to look into speeding up that code. should look into a SecureString implementation too.

just found that Sebastien Pouliot has a really old post about the new classes in System.Security.Cryptography for .NET v2. he has been heading the security implementation for Mono, along with their WSE stack.

i decided to go ahead and implement the algorithms that Whidbey B1 (desktop) adds to System.Security.Cryptography. just finished up RIPEMD160, HMACMD5, HMACSHA256, HMACSHA384, HMACSHA512, and HMACRIPEMD160. will try to add ProtectedData, ProtectedMemory, and Rfc2898DeriveBytes tomorrow ... if i get those working, then that will give the CFv1 as much crypto as .NETv2

just wanted to say that i do not care

just took my 1st trial run. the install hiccuped on the device emulator, but worked on the 2nd attempt. MSDN documentation would only install if i went and killed some of the lesser 'msiexec' processes that were spawned, otherwise it just hung. just did a real quick PPC app. 1st thing i noticed was that the emulator connected to the internet without me having to change in settings. this will make developing mobile web apps easier. also looked at the MS System.Security.Cryptography namespace for CF. the included algs are: DES, DSA, MD5, RC2, Rijndael, RSA, TripleDES, SHA1. that leaves out: MACTripleDES, HMACSHA1, PasswordDeriveBytes, SHA256, SHA384, and SHA256. while on crypto, the full framework adds the algs: HMACMD5, HMACRIPEMD160, HMACSHA256, HMACSHA384, HMACSHA512, Rfc2898DeriveBytes, and RIPEMD160.

: a practical introduction to AI (published 1989) is the title of the book i just finished (while installing Whidbey earlier). its another good book that gives an outline of the different AI techniques. it points out why one process might be chosen over another, due to advantages or disadvantages. it also does a good job of comparing and contrasting a school of thought to how it is believed humans actually think. the only negative of this book is that the code examples were in POP-11. anybody know if POP-11 has a .NET compiler? (i'm not serious) the code for each chapter builds upon previous chapters to ultimately produce an automated tour guide app. the project was well chosen to demonstate the different ideas.

one too many lil john videos
carwhistle.wmv (3853 KB)

mean commercial
fur_is_dead_master.mpg (6681 KB)

1st) my WSE implementation for the Compact Framework made it as a late entry into yesterday's OpenNETCF SDF 1.2 Release. the bits have been updated for WSE 2.0 Release to support WS-Addressing, WS-Attachments, WS-Security, and WS-SecureConversation. also have a corresponding article with client side code to get online. it can be used to call every WSE sample except Kerberos
2nd) just as soon as i get WSE 2.0 support out the door, Hervey announces the coming of WSE 2.0 Service Pack 1. dont see anything breaking within the change log ... but i would love to test an alpha version before this is released! actually, i've been failing at getting ANY inside track to what MS is doing with Web Services ... so i'm not holding my breath :( also, none of the pain points i've had with WSE 2.0's handling of non-default encryption algorithms show up in the SP1 list ... but that is ok, because device developers enjoy pain :)
3rd) actually, i'll relieve most of that pain shortly. i've got an extended crypto library for CF now that supports AES that has already been integrated with the WSE bits in OpenNETCF. meaning once i put the crypto code out there, then a device can use this additional library to call WSE 2.0 using its default encryption algorithms ... including passing Entropy for WS-SecureConversation. then the only changes that the samples need to have done are getting rid of the use of 'localhost'
4th) there is alot more work to be done ...

for work
hn_gggbaghdad.wmv (1271 KB)

gotcheats.avi (4142 KB)

bad
bechau1.wmv (6100 KB)

Chris Tacke and Neil Cowburn of OpenNETCF Consulting are sponsoring a coding competition. one of the categories for entry happens to be 'security'. would love to see some entries that used the OpenNETCF.Security.Cryptography namespace :) but security is much more than cryptography ... how about some biometrics, or secure apps! BTW i'm not an OpenNETCF board member (so could compete) ... but will not be. cannot wait to see what you come up with

moved into the new house. we immediately had electricity and water. even a phone line if we did not opt for cell phones instead (no DSL in this area yet). yet it took a week to get cable (and internet) setup. dont care about TV (except cartoon network), but i'm entirely crippled without internet. i communicate more through email and messenger then i do over the phone. we did not even get a phone book, since we did not get a phone line, so how was i supposed to have any idea where to find food and such? regardless ... i'm back online now!

fixed the bug i was having with AES128 yesterday. ended up i was doing CBC padding where i needed to be doing ECB. with that fix, i was able to call the WSE 2.0 samples without having to modify the server side! also, this let me call the WS-SecureConversation sample with Entropy (since the STS cannot be configured to do TripleDES Entropy). the Entropy sample works by generating a random key on the client and passing it as Entroy to the STS, encrypted with RSA. the STS then generates its own random key, and passes that back, AES128 key wrapped with the key passed in the request. the client can decrypt that key wrap with the key it originally passed to the server. the client then does P_SHA1(requestEntropy, responseEntropy) to derive the secure conversation key to securely call the web service.

got my 1st recruiter call last week for WS-Security ... this is basically how it went. recruiter: do you know WS-Security? me: yes! recruiter: how well? me: i basically wrote it from scratch. recruiter: this project is in java ... me: [hangs up]. the last part was embellished :) wonder if a recruiter will ever contact me for .NET WSE?

got some answers back: 1) the Mono class libs do fall under MIT X11 (runtime and C# compiler are GPL or LGPL). even though i dont consider a single class 'substantial', the copyright does need to be included in a single class that use in my own project. this isnt a problem because the copyright is very open. 2) not all Mono authors transferred their copyright. this makes me think that you could contact those authors directly for the code, and ignore the Mono license if they let you use it. aside, a quick search through the code shows that most of the source files are marked differently. even if you find the same class twice, 1 instance will include the MIT X11 license, while the 2nd version will not. i did come across a Novell copyright in some of the files too.

got the CF AES port partically working WSE 2.0 now. the actual encryption and decryption of aes128 with CBC padding is working great. now i am working on the key wrap algorithm. my test harness shows that it is mostly working, except i am having some difficulty with it being in EBC mode (one unit test on the device is giving diff results than the desktop). the pseudo-code for implementing the key wrap algorithm from the Xml-Encryption specification was a real PITA too. instead of pseudo-code, they should provide an actual implementation in some realistic language. i dont care what language ... just that it will compile. then i could check that code out to see where mine breaks. at least they do provide some precondition and postcondition values.

from the Mono Handbook, the class libs fall under the MIT X11 license. this is a very open license that says the copyright must be included in a 'substantial' piece of the software. if i were to grab a single class from the class libs of mono, which i dont consider 'substantial', do i even have to include that copyright? am i getting this entirely wrong?

not able to get to the mono site right now ... but i have a question: did the authors of the source for mono transfer over their copyright? it is not a work-for-hire, but did they sign over their copyright? which would now be owned by Novell. from what i remember (from beta downloads), each source file had a copyright from the author, and some even had multiple copyrights from the individual authors that worked on it over time. if an author still had copyright, couldn't i just ask them for a specific piece of code that i wanted, and ignore the whole GPL nonsense?

from a couple posts down ... i brought the AES C# code over to the compact framework. also the C code with a C# wrapper. both tested out to be compatible with the full framework Rijndael implementation. the code wrapping the C lib is about 5 times faster than the pure managed code. the codeproject article also had a managed SHA256 implementation, and some PasswordDerivedBytes updates. now i need to extend the CF WSE bits to be able to use this enhanced crypto library.

the 1st movie was better than expected. 2nd was worse than i thought the 1st was going to be. now for the soundtracks: the 1st soundtrack was split between rap and techno. something like 6 rap songs and then 6 techno songs. the 2nd soundtrack merged rap and techno. used techno DJs with rappers for lyrics.
Blade_Trinity_Trailer1_700_dl.wmv (7158 KB)

when it starts up the very 1st time, it asks what runtime i want to target. is there a way to reset that ... other then deleting the app.exe.config file? what i've been doing is copying the .exe a couple different times, and just executing a specific copied version depending on if i'm currently targeting the desktop or a device with the compact framework. would love to just retarget in the same instance.

i've got 1 test to go til MCSD'ness. right now i've got MCAD, which means i've taken one test to be MCP. what is the etiquette? am i supposed to list MCSD only? i usually see people list all of them in descending order. e.g. MCSD, MCAD, MCP ... which seems like padding. and are people adding the .NET for MCSD.NET and MCAD.NET? just wondering what the cool people are doing

just what i had been looking for: A CBC Stream Cipher in C# (With wrappers for two open source AES implementations in C# and C). this takes about 2 seconds to port to the Compact Framework, and my initial testing shows that it is compatible with the WSE 2.0 AES implementation. excellent!

been waiting forever for The Prodigy to release the next album. got a preview of it yesterday. not as mainstream as 'fat of the land', but not quite 'music for the jilted generation'. the 1st single to be released is supposed to be 'girls' ... which is probably my least favorite track. wish that 'babys got a temper' would have been a bonus track

success! just called the DerivedKeyToken sample. it works by 1st requesting a SecurityContextToken from a token issuer. as mentioned in an earlier post, the client cannot pass Entropy, so the token issuer generates the key and returns it encrypted. from that key, the client derives 2 new keys: 1 for signing, and 1 for encrypting. then it signs and encrypts the request to the web service using each derived key respectively. using the info the client passes, the web service can derive the same keys to verify and decrypt the request.

that milestone basically completes the update of the CF bits for WSE 2.0! i need to clean up the code, do regression testing, and a write-up before releasing the code

doing a re-install of vs.net 2003 fixed everything. just cost me a significant amount of hours. i've seen a number of postings in which peoples device dev environments have gotten hosed up. this is my 2nd time ... full re-install required both times. just wish there was an apparent reason for it having messed up? i.e. it was working great, and i was following my standard dev patterns, and then it just stopped working?

was about to debug calling the WSE DerivedKeySample from CF on my Pocket PC. ... application deploys ... hourglass comes up ... app never starts. crap! repeat debug steps ... and get same result. crap! create a simple HelloWorld test app ... crap! soft reset device ... crap! install latest CF service pack ... crap! install latest ActiveSync on desktop ... crap! install win ce utilities for vs.net. DelCryptoKeys ... crap! SdAuthUtility ... crap! restart computer ... crap! hard reset device ... crap! re-installing vs.net 2003 now ...

... and Sam too

helping the TS buy a house. the home loan people have successfully pissed me off by asking me to lie on a form. i have serious issues with dishonesty, but will submit because this isn't really my deal. would handle differently otherwise: 1st, thing i would do is require them to show me legal documentation of why they want me to lie. they have not provided a valid reason yet. 2nd, i would start going up the chain of managers until somebody could get me out of lying. 3rd, i would find another lender ... which would probably get them to backpedal real quick. the form has already been signed by me, they just need to change one word, so i will turn a blind eye. wont change it myself out of principal.

have got back to working on getting the CF bits able to call WSE 2.0. the good news is that i've been able to call the WS-SecureConversation sample, but not with Entropy. the bad news is that the encryption algorithm for Entropy is not configurable server-side (by-design) and that it uses kw-aes128 by default. since AES does not currently exist in the CryptoApi for SP 2003 or PPC 2003, then this feature is out. my hope was that i could change the entropy encryption to kw-tripledes ... and then we'd be golden. i'm not holding my breath to hear the design decision of why this is not configurable :)

housejump.wmv (4869 KB)

in texas you only play soccer until you are old enough to play football
RUUD3.wmv (4817 KB)

young minds are easier to corrupt
jihadkids2.wmv (563 KB)

just making sure you read Michael Yuan's latest post: Nokia's big plan to beat .Net CF in mobile Web services. guess this is their response to the Devices Profile for Web Services.

sub000984_1_best.wmv (9865 KB)

this kid makes me laugh my but off. the spider has found tons of videos from him. high quality
sushi.wmv (2446 KB)

test with a better thumbnail
iv.mpg (1559 KB)

this is the 1st test of a new video link blogging setup
bmwRally.wmv (966 KB)

minding my own business driving to the gym ... and then all of a sudden i'm assaulted by a graphic ad campaign against abortion. it was one of those big trucks with advertising on all sides. they would drive about 5 miles, make a U-turn, drive back, U-turn, repeat. after i got done working out, they were still circling. the funny thing is that they had a security escort following them. have you ever heard of a pro-choice group attacking an anti-choice (aka pro-life) group? me neither. i would like to see a pro-choice campaign showing graphic pictures of doctors that have been killed by pro-lifers (oxymoron).

listened to the book Oryx and Crake on the drive to WI. it was a rather disturbing look at what genetic sciences might come to in the future. one part of the book that made me chuckle was a refrigerator magnet along the lines of: "i am, therefore i spam". classic.

just finished another award winning piece of literature from ireland. have 2 problems with them in general. 1) way too depressing. 2) if it is in audio book format, then it is usually read by someone with a heavy irish accent; making it difficult to understand what is being said.

drove past 2 white kids (approximately 20 years old) picketing 'fahrenheit 911' playing at a local movie theater. missed what the 1st sign said, but the 2nd read 'fahrenheit 911: communist propaganda'. not sure if i'll see the movie because it sounds just like the book 'dude where's my country'. still need to see 'dodgeball' instead. didnt have my camera phone with me at that time, and they were gone by the time i got back :(

this one was near the rock stage

made it to Milwaukee last night after 13 hours of driving. made 2 rest stops in Ohio, with stories following:

stopped at the 1st one to eat. went outside to the patio where it was a little windy. a big gust blew my trash away which i retrieved and threw away. sat back down and an older man came up and said 'I admire you.' ran through my head in what i could be admired for ... saving the environment, relative youth, rugged good looks :) my reply, 'what?' he says, 'you were praying, weren't you?' he must of thought that i was blessing my food when i put my head down to rub my eyes, because they were road weary. my answer, 'hell no.' then i left

second stop was for ice cream! older guy in front of me turns around and says 'remember about saying the pledge of allegience in schools?' all i could think about was how i remember not to talk to strangers. my answer, 'yes.' old man shows me his money, 'look right here, it says: in god we trust. first time i noticed that.' had already noticed this, and replied, 'we better reprint all the money.' then i proceeded to pay with my secular debit card, 'in visa we trust'

from a news.com article: Multiple standards keep customers from taking the Web services plunge. 'multiple standards' sounds like an oxymoron to me? regardless, it takes common sense to figure out the standards mess. if it says Microsoft and IBM on the standard, then it is gold, bet your business on it. if it says Microsoft and others (without IBM), then look for a competing spec my IBM and others. if there is no competing spec then proceed forward with caution. if there is a competing spec, then wait. if you see a spec without Microsoft or IBM behind it, then throw it in the trash. also pissed that the Yahoo! Groups "WS-Discovery-Workshops" discussion group requires me to sign and fax a Workshop Feedback Agreement. hate that legal is involved and that it requires killing trees. the only WS-good news i have is Christian Weyer, John Bristowe and/or Interop Warriors have put together Plumbwork Orange. this will be used to test out WS-* implementations. needed since time between WSE drops has been substantial. don't know if i'll contribute to the workspace, but i'll definitely be using it to test interop with the CF bits. need to run the WS-Eventing stuff against it now

i leave boston tomorrow. so i will have to see Phil, Sam, Jason, and Greg in the blogosphere. leaving because i found boston way too racist ... just kidding (barry bonds joke). didnt run across any of that at all. am heading back to milwaukee for an indefinite amount of time where the TS will be domesticating me. if the blog starts falling off, feel free to making whipping sounds. taking a week off for summerfest, and then i'll be looking for work. any reality TV shows auditioning ... i'm up for that too? also will push to get the updated CF bits for WSE 2.0 ready for a preliminary drop.

this officially ends the 'brawn' series of posts. figured i would end with a bang ...

working out has sexual benefits. the 1st is getting in shape so that others will want to have sex with you. more self confidence, etc, etc. 2nd, it can help you perform better. i think the best exercise is squat, for multiple reasons. it is a whole body exercise that works out your entire body. this means increased blood flow and cardio benefits. think increased blood flow is what viagra and derivatives do? it also makes your body release testosterone. something about crushing your body with weights will do that! if you notice in your gym, the guys that squat will typically be the strongest on bench press too. but you wont see many people squatting, because it is a very difficult lift. you'll want to go parallel to get the most benefits. if you dont squat very low, you will mostly use your quadriceps, which has little benefit. if you get low, your quads are taken out of the picture and the hamstrings kick in. this is where you get thrusting power :) actually, the real benefit is kegel strength. could go into gory detail, but i'll spare you ... and you can do your own google search to find the benefits of having a strong kegel. going below parallel is really tough, but is hard on the knees. squat is the best, but the other leg exercises will help too. other good ones are the leg abductor and adductor machines. you know what i'm talking about. these are the machines that have a girl lifting weights with her legs spread-eagled. usually a scene in old cheesy late night movies. also called the pussy tighteners. for guys, they are kegel strengthening machines. most gyms offer these machines so that you can do these exercises with your knees bent. this feels much more masculine than the straight legged variety. you just have to get past the misconception of them being girl-only machines. the funny thing is that most guys that lift do not workout their legs at all, and most girls only workout their legs. total gym segregation. so that is another reason for guys to workout their legs ... to be in the same vicinity as the girls in the 1st place.

Bill reminds me that i used to be more fun. dont know why i stopped writing silly programs? definitely need to take some new Whidbey tech and totally misuse it

no ... i dont have any gmail invitations ... back off. this news.com article states that Google (Gmail) will not display ads on e-mail messages with words related to sex, guns, drugs and other topics it considers off limits. this means that i wont be seeing any ads at all! :)

this is what i hear most often.

1) your site looks like crap in browser X? A: i only test in IE. actually, i want to rewrite my website, but its not at the top of my list. probably when ASP.NET 2 gets closer. even then, i'll still only test in IE ... so no promises
2) you dont allow comments? A: i sort of do. there is small comment(s) link where you can leave short messages. it is just a running list and is not tied to a particular blog entry. thats all the functionality i want to provide for right now
3) use CAPITAL letters? A: no. why is that so essential. i use periods to end sentences, why do you need a capital letter to see that a new sentence is starting. that would be redundant
4) have you heard of the ENTER key? A: yes. but if i used the ENTER key then my right pinky would get much stronger than my left, since i wont use CAPS
5) provide categories so i can filter out the offensive stuff? A: i do need to add that feature
6) provide categories so i can filter out the nerdy stuff? A: see #5
7) your permalinks are broken? A: yes, its a problem with my XSLT. will fix that when i get access to my server
8) why no ping/trackbacks? A: might implement that as well. then i could see how very few people have the balls to link here. that is, after i fix the permalinks
9) why dont you just use a standard blog engine? A: to keep some web development skills
10) why arent you linked from MSDN? A: dont you see WARNING at the top of the blog?
11) why does my company's proxy block your site? A: see # 10

just finished reading 'MIND Children: The Future of Robot and Human Intelligence', circa 1988. although dated, this was a really fun read because the future it envisions still has not happened. the author did bring up nanotechnolgy and the mars rovers as being promising :) overall, it has an evolutionary theme where the machines we develop might be our next step. i.e. there were some Borg and Matrix themes. started out contrasting robotics and AI. then did some computations to try and figure out when a computer would be powerful enough to model the human brain as a general purpose AI. next, talked about possible relationship scenarios between man and machines. finally, wrapped it up with viruses, the big bang, and fourier transforms. this stuff is great for making you want to think BIG thoughts

got a long drive coming up, so i was grabbing some audio books for the trip. have a number of audible.com books so i downloaded the latest software. it tried to load some bits on my PPC 2003 device, and gave me the error below. what was real ugly is that i had to kill the process to get it to go away. also wish that the audible software would create mp3 CDs. couple google searches turned up a shareware program called GoldWave. it let me directly convert the audible files to mp3. high quality. now i'll be able to make the trip without having to swap CDs every 80 minutes

visual studio team ... blah blah. all i want to know is if it is going to support the compact framework? i'm talking unit testing and performance analysis on a device. otherwise i could care less

following StronglyTyped's lead. my results:

You demand a free and unattached life for yourself that allows you to determine your own course. You have an artistic bent in your work or leisure activities. Your urge for freedom sometimes causes you to do exactly the opposite of what expected of you. Your lifestyle is highly individualistic. You would never blindly imitate what is "in"; on the contrary, you seek to live according to your own ideas and convictions, even if this means swimming against the tide.

the secondary pictures i liked fit pretty well too. better yet, the pictures i did not care for, did not fit. the best match i've ever found though has been INTJ. i have to substitute 'questionable GNC supplements' for 'substance abuse' though

what does the 70 mean? is this the seventh series of certification exams? was there a 69 series? ... wink wink. switching topics: i second guessed myself alot during this exam. the case studies were definitely contradictory in places. in real life, i would have asked somebody to clarify ... instead of just making my own decision. at one point, i could have done the app as a web app or a windows form. a couple of questions were dependent on which i chose. i took the option as answering one for the web app scenario and the other as a win form scenario ... to only miss one, instead of possibly missing two. no glory in that! the other thing about these test is i know TOO much. sorry for the ego on that last sentence ... but let me explain. reading blogs and newsgroups have taught me many tricks and exceptional cases you can use to get around certain problems or limitations of .NET. for a small number of questions there would be 2 answers that were both correct. with those 2 answers, i would then have to filter it down to the mainstream answer that the test would expect, and throw out the more esoteric workaround.

just know that i passed. it did not tell me my score ... so i'll assume i aced it :) this test was definitely different than the other exams. glad that Jason Haley tipped me off on How to pass 70-300. i prepared by reading the MsPress book, which gave me the necessary jargon and frame of reference, but was somewhat laughable to read. searching around it looked like the ExamCram book might have been better? also getting a look at what case study questions were going to look like was essential. i recommend running through the Case study-based test demo and the free Transcender demo for the test. that puts me at 1 away from MCSD. will probably go with the new Security elective. anybody know if Biztalk has a beta going on?

for the sports fans ... i used to play little league baseball with the fan that took the ball from the little kid

you can read a couple blogs down about me getting blasted as a spoon fed .NET flunkie in the platformsdk.security newsgroup by a C++ bigot. if you follow that link you'll see that i sent off a cursory reply. by cursory ... i mean i cussed. the thing to note is that the message does not show up in Outlook Express. but the message does show up in groups.google.com. its like google has more of a direct link into the newsgroup than OE does? else the google spider crawled over the message before it was deleted. the groups spider has something like an 8 hour delay, so that is unlikely. maybe OE is filtering messages that have cuss words in them ... not likely. this one has been bugging me. i need to know when and where my cuss words will appear. fuck ... just testing :)

this is about my 5 month mark of blogging. this is also about my 6 month mark of reading blogs (had been ignoring them). here is my current take.

LIKE: 1) that i get to hear directly from the source of people that are developing the technology. if i hear it from somebody else like me, then i'm like 'what the hell do they know? spares.' 2) when blogs have a personal side with opinions. being funny is also a plus. if i can learn something technical and laugh ... win squared. 3) great for up to the second info, because i cannot wait until tomorrow to find out that product X is released. 4) newsgator outlook integration. chocked full of heady goodness. 5) comments and trackbacks are fun to peruse. 6) that anybody can blog without being technical. anything to take some control away from big media is good. 7) quick medium that i can write to and publish immediately. for when i dont want to think about what i'm about to say and how it will damage my professional career. 8) moblogging. because i need some new ideas to implement with the compact framework. 9) finally, its a great way to keep fresh content on my homepage.

HATE: -1) the whole RSS vs ATOM thing. throw them in a standards body to be merged and move on. -2) trying to find old blogs i need to refer to again. worse than searching my harddisk. -3) hearing the same information from 10+ different people. there should be a heirarchy of blogs. i.e. if don box blogs something, then i know that i dont need to say shit, because you already heard it from him. -4) when people change their blog titles and newsgator makes a new folder. -5) when people update their blog feed (or individual blog) and newsgator makes a copy of the message(s), instead of just overwriting. i know that is configurable, but i dont do it. -6) that people bitch about getting blog spam, but use the same blog engine as everybody else. if you want unspammed comments, then write your own comments page. also stop answering questions in the newsgroup about how to programatically do HTTP POSTs to a Url. -7) when individuals that aggregate blogs think they are getting copied all the time. not everbody subscribes to you, or maybe they found the info independently and just happened to miss your post. this is in conflict with my blog heirarchy complaint above. -8) aggregate feeds. my blog is too offensive, so i cant join in. but i would join an aggregate feed of offensive blogs. -9) dave ... could not resist.

OTHER: what about the other communication channels: 1) miss the ListServs. there was no better way to ask a question and have it answered immediately. i try asking questions on my blog, but nobody answers because they already unsubscribed to keep their virgin minds pure of bad thoughts. 2) still use newsgroups alot. threaded messages are a good thing. also groups.google.com for searching. figured we would have blogs.google.com by now? 3) articles are still good. want them to have more permanent links off of my site, and i usually spend alot more time putting them together. 4) still read books. they help to organize all the random bits of data. outlines are goodness too. 5) still use email. because contrary to what you might believe, i keep some communication private for the intended recipient and anybody sniffing packets on a web proxy.

PEOPLE: there are new roles in the industry: 1) now people that dont have the technical ability or character defects to keep a website updated can have a voice. e.g. your mother. i just wanted to work 'your mother' into this somehow. done. 2) now people that dont make content can just aggregate it and give their own opinions. e.g. Scoble. i'm going to start calling these people developer analysts. glad i dont have the word anal in my job title. 3) people can just be funny. e.g. Rory. i dont care if you are technical, if you're funny, then i'll subscribe. cant do that in the newsgroups i visit. 4) the person that is the direct source of the info can let you know what is going on. e.g. Box. without the spout then we would miss alot of gems between books.

SUMMARY: ignore all of the above ... blog for google juice alone. you'll have your blog which google will crawl over. then you will link to others from your blog and your blogroll. they will be kind and link back. then you will comment in their blog and leave your URL for more juice. then they will implement trackbacks and if you point to them then the link will show up on their trackback. you'll get into an aggregate feed for more publication ... plus blogrolls. the other communication channels cant touch that level of viral inbreeding.

got it to work by using the extended TripleDES classes. not entirely calling the sample as intended, because i'm not currently passing 'Entropy' to the SecurityContextToken issuer. the scenario is that the client passes some random bits to the server, and the server returns some random bits. by combining those bits, they can both generate the same key to communicate with. want to get that working, plus the DerivedKeyToken sample. going to have to refactor some code before i can get all that working

Andy Sjostrom's DataSet Server CE article has got me thinking of a SILLY idea. my Pocket PC can do SQL CE, but my Smartphone cannot; and i usually carry both devices. what about exposing a service on the Pocket PC, wrapping SQL CE, for my Smartphone to dump data to. hmm ... great article Andy

figured it out while pretending to listen to the TS on the phone :) the problem was that TripleDESKeyWrap uses the TripleDES algorithm with no padding. so that if i encrypt 16 bytes, then it returns 16 bytes. the .NET Framework provides an enum to turn off padding, while the CryptoAPI does not. to make the CryptoApi compatible while encrypting you just toss out the padding bytes. e.g. you encrypt 16 byes, it returns 24, and you just throw away the last 8 bytes. decrypting was not obvious to me, and led to the blog below. so i have the key, and the last encrypted block. to get the padding that was chopped off you set the last encrypted block as the IV. then you encrypt the padding bytes. for 8 bytes of padding this is: 0x08080808080808. encrypt that with the CryptoApi, and it returns 16 bytes with the 1st 8 bytes being the padding that i needed to derive. append that to the original cipher data, pass that to the CryptoApi to decrypt, and you've got a compatible implementation of TripleDES with no padding. now i can finish the TripleDESKeyWrap implementation, and then continue on with WS-SecureConversation

of course i deserve this ... but check out the love i got from a native coder.

Why is it that so many of the .NET flunkies don't have good programming common sense or the will to try to code themselves out of a corner?

had not been able to get the WS-SecureConversation sample to run at home. have been able to run it on a different machine, and the only real difference i could think of was i have Whidbey installed at home. start digging around, and check out the new tab i found. its in Contol Panels-Admin Tools-IIS-[VIRTUAL_DIR]-Properties-ASP.NET tab. set it to .NET 1.1 and the code now runs! oh yeah, somebody should look into why the sample wont run under Whidbey May.

on the web services security specifications page of the MSDN web services developer center, you'll find this beauty:

Note If you are implementing the WS-Security specification, the Web Services Security Addendum, or the WS-Security Profile for XML-based Tokens, please print, sign, and return the Web Services Security license at this page to Microsoft Corporation.

just so happens i have implemented a chunk of these specifications. so when i go to read the license, i get this:

Whereas, Company (brains-N-brawn LLC) wants a license from Microsoft to implement WS-Security, and whereas Company understands and acknowledes that licenses from IBM, VeriSign, and/or other third-parties may also be required to implement WS-Security.

err, umm ... i'm supposed to do this crap for VeriSign and IBM too! hell no. isn't OASIS supposed to protect me from this crap?

WSE policy files will let you specify how to secure your web services without mucking with the code. concentrate on the business logic ... blah blah. but suppose i am hosting my web service at a 3rd party site. seems like somebody could just replace the policy file with a new one that is less secure. while if i put the security code mixed in with the business code, then it gets the additional security protection provided to .NET assemblies. am i wrong about that being a trade-off? what about signing policy files themselves? maybe i sign the assembly with my private key and i also sign policy file modifications with the same private key. then the assembly can use the public key to verify the policy file. if somebody replaces it with their own file, then WSE will see that it is not properly signed and fail. if i need to modify it myself, then i just sign the new policy file and the running assembly will be able to pick it up. i really have no knowledge of WS-Policy ... basically just making this up ... somebody tell me i'm wrong

msmobiles.com reports a ROM Upgrade for i-mate Smartphone2 available. the i-Mate Smartphone2 is one of those elusive Smartphone 2003 devices that has the Compact Framework in ROM. which brings up the point of this blog ... why isnt the Compact Framework SP2 in the fix? my hope was that Smartphone devices would be easier to develop to because they are always connected and service providers could push down the latest stuff. on the contrary, the pattern now looks like you will have to upgrade hardware to have the latest software

the CF WSE 2.0 bits are coming along. most everything is working except WS-SecureConversation. after i get that, then i'll probably clean up the code and release it to the wild. still alot more that could be done to it, but people are starting to actually ask for it. also studying for the 70-300 exam. want to try and knock that out this weekend. reading the MsPress study guide right now ... really boring. a little scared of this test because i know 3 people that did not pass the 1st time, although they aced it the 2nd time around; and passed all the other exams easily. finally, i've got 1 more week in Boston. contract is over and i'll be heading back to Milwaukee

weighed in over 225 today. thats about 5 pounds since upping calorie and protein consumption. gaining weight is basically the easiest way to get stronger; except now i'm gaining strength too fast. my right elbow and shoulder are hurting a bit, because the joints are lagging behind. will have to redirect workouts to other areas and give those joints more time to recover. also going to slightly reduce calories to try and level out at this weight.

(for causticPhil sake) this one has been bugging me lately about the subway. everybody is in a big hurry to get to (and away from) the subway station. the collective walking pace speeds up the closer you get. so everybody is real important and has to get to where they are going ... Now! you can even see this when getting off the train. people start milling around the door so that they can get out 1st. the funny thing is that they will dash out the door and immediately form a line to the escalator, even though the stairs right next to them are mostly unused. now i see that the escalator would be faster if everybody was walking up the escalator ... but that has not happened yet. somebody always stops, effectively making the stairs the faster means of transportation, on which i pass them all. the parking garage is even more odd. there are stairs you have to take up and escalators that are going down. but in the evening the escalators are turned off, so you can walk up them. what bugs me here is that everybody is crowded on the stairs, and nobody is walking up the escalator. my guess is that they are scared it will turn back on and they will either have to try and run up it against the grain or retreat back down to take the stairs? or they realized how stupid they were for waiting in line for the escalator in the morning, and they vowed to take the stairs; but now the stairs are the wrong choice, so they vow to take the escalator in the morning ... a vicious cycle of nonsense

think i woke up with an A-HA moment! started to implement the kw-tripledes logic, but the return from the web service was only 32 bytes of data. if i was following the logic correctly, then this would have only left me with 16 bytes of key data, while 3DES needs 24. so i hope what is happening is that i told the web service to use 3DES for key wrapping, but i also need to tell it to use 3DES for secure conversation keys; and that right now it is returning an AES128 key (16 bytes). my original assumption, which i hope is wrong, was that kw-tripledes would be used to wrap 3DES keys only. if that is the case, then it would be helpful for the 'EncryptedKey' element to specify what flavor of session key it was wrapping; especially since the size of the byte arrays for TripleDES and AES192 would be the same ... at 40 bytes.

got to take a quick look at the WS-SecureConversation changes. the most obvious change was the addition of the 'Entropy' element. this lets the client pass some random data to the server, which the server can then add its own randomness to, to create the shared key. kind of cool, albeit optional. ran the sample without passing Entropy and it returned an EncryptedKey as expected. what was interesting is that the security token the client passed was a UsernameToken, so asymmetric encryption could not be done. instead the server generated a key from the password and returned the EncryptedKey as kw-tripledes (Key Wrapped Triple DES). looking at the spec for kw-tripledes produces this gem: Decrypt the cipher text with TRIPLEDES in CBC mode using the KEK and an initialization vector (IV) of 0x4adda22c79e82105. where in the world did the magic IV come from? ... this is going to be tougher than i thought

there is signage on the subway for the 'gay men domestic violence project'. says something to the effect that 1:4 gay men are abused. that number seems high (or low) to me. is the percentage HIGH: haven't they been fighting for same sex marriages? in fact, its now legal in Boston, where the signs are in the 1st place. if 25% of their relationships are abusive, why dont they break up? all i can figure is that they are including prison rape in those numbers. and if you are a guy and want to abuse somebody, wouldn't a woman be much easier? [NOTE that is a joke. tasteless ... yet still a joke. get over it. i would NEVER hit a woman, and would beat the crap out of guys that do.] back on topic. is the percentage LOW: at least 1:2 of them are taking it up the ass ... which i would definitely consider abuse. related; how come i dont see signs for lesbian violence? would pay good money to see that

Main Entry: caustic
Function: adjective
1 : capable of destroying or eating away by chemical action : CORROSIVE
2 : marked by incisive sarcasm
3 : relating to or being the surface or curve of a caustic
-Copyright 2004 Merriam-Webster, Incorporated

1st. your welcome.
2nd. Thank You!
3rd. why is your blogroll alphabetical, but i'm at the bottom?
4th. less capital letters please :)

SBC posts that there is now a security elective for MCAD/MCSD. got MCAD, and had been complaining that i would not finish MCSD until they came up with some newer electives ... was hoping i could ride that excuse for a while :) time to get back to studying :(

Softwaremaker is doing some great WSE blogging. also tearing up the ms.pub.net.fx.ws.enhancements newsgroup with me. got a number of open items in that newsgroup that i hope we get answered. also hoping he gets a stand-alone SecureContextToken service working. think that i'll need that gem to get the Kerberos sample to work.

there was a great post recently along the lines of there being too many WS-* specs, but that will eventually be hidden from us. want to give credit, but i cant find that post again ... one thing i hate about blogs. thought google was going to fix this after they bought blogger? back on topic ... i like all the WS-* specs and trying to make the CF bits interop. its the right level of pain for my masochistic side. i'll give you an overview of what this involves: the specifications are key, and i read them, but they allow room for differences in implementations. at that point the WSE implementation supercedes the specs, and i write the code to do what WSE does. mainly this involves the shaping of XML. the XML has to be very precise. namespaces have to be correct, the order of elements might matter, the xpath of the element. add in cryptography and certain operations have to happen before others. so the interop involves me looking at what XML the server expects, and comparing it to what the CF clients are sending out. this is where i am so happy that XML is text based and not some binary format ... except ... crypto brings back the binary format. when trying to get crypto to interop, i am constantly looking at byte arrays. have helper methods to compare: -look for any match with different sized arrays, if so, then an IV might be appended. -look for a match but the byte arrays might be reversed. -look for a match with values off by 1, might indicate salting. etc... as a friend put it, you have a set of switches in front of you. if you flip a switch and the byte arrays dont match, you try another switch. with some trial and error of different permutations ... match!

love futurama, but thats not the topic. nor is killing yourself. the topic is the mixing of various fountain drinks. this is what i do at fast food places. blend just about everything: diet soda, root beer, sprite, tea, powerade, fruit juice, whatever ... all into 1 new drink. if there is a big selection, i'll get like 8 different ones in. its like having a new drink every time! the thing is not many people do this. also today was my cheat day, so i ate at 2 different fast food places to satisfy my craving for fatty foods. fast food place #1 had the fountrain drinks out in the open. the problem was they wanted to know which fountain drink i was going to get. i told them they didnt need to know what i was going to choose, and then gave them some data mining and privacy nonsense to try and confuse them. they said they needed to know so that they would know when to order more supplies. had to proceed to explain to them about the suicide ... and they finally gave up the cup with a look of disgust. fast food place #2 had the fountain drinks behind the counter. in this situation i usually just give up and pick 1; except i'm getting more eccentric in my old age. told the guy i wanted everything except plain coke. after the initial shock wore off, he proceeded to make my drink. the funny thing is it took him a long time because he tried to get equal parts of everything! also the drink machine was made to auto fill the cup to the proper amount, so he had to override that. somtimes it is really hard to not conform

lower body strength is starting to come back. just did 810 for 5 reps on a slider. still got a ways to go though because i was doing just over 1000 last year and 1250 (for reps) is my personal best. want to get back to the half a ton range. its just fun to say half a ton :) now for the finer points of leg press. there are basically 2 types of machines: pivot and sliding. the pivots are the easiest and you'll be able to do about 25% more weight on them. the further you get your feet from the pivot point gives you more leverage and makes it easier too. the sliding kind varies based on the platform where you place your feet. on these, its a little easier to get your feet further up too. if the platform has a bend in it, then putting your feet below the bend will be substantially harder. sometimes the seats are adjustable as well. the further the seat is up makes it easier. this is mostly because your range of motion will be reduced by your knees hitting your chest. its better to put the seat all the way down and get a full range of motion with lower weight. leg press is good because it is self spotting. if you get in trouble you can push on your knees to help out. now to make you cool. if you lift heavy weights, then dont use the 5 or 35 lb plates. all you really need is 10s, 25s and 45s. putting 315 on the bar and then adding 5 lb plates on each side looks stupid. and when you do heavy weights on machines, you stop worrying about the weight of the machine itself. so when i say 810, i dont count the 25 to 45 lbs of the leg press carriage. for one thing, they vary from machine to machine, also nobody really knows how much the carriage really weighs. its not common knowledge like a barbell weighing 45 lbs. better yet you say the number of plates on each side of the machine. e.g. 810 is '9 plates'. 9 45 lb plates on each side. this helps because weightlifters are not the brightest. if i'm doing alot, sometimes a person will come up and try to figure out how much weight it is ... and it will take them way too long. the worst is when i guy when to the front desk to get a calculator.

got this one working in about 5 minutes. reused the same web reference proxy i had created for calling the CustomBinarySecurityToken sample, all that was needed was to change the Url to point to the Xml flavor. then it just magically worked. this improves my confidence that the WS-SecureConversation sample will be easily called once i can get the server side pieces working

the movie 'bound' is one showtime, and a diff lesbian show is on another showtime channel. hope lesbians start getting more airtime than the queer eye guys. hadnt thought of that ... female eye for the straight girl; i'd watch that too. about time media starts promoting it.

was browsing barnes and noble and ran across The Microsoft Platform Ahead by David Platt. have not read it yet (except for the part i reviewed), but will because i love his writing style. he does a great job of getting across what is important, while being funny. now for the hidden agenda ... if you flip to the preface, my name is listed as a tech reviewer. kind of cool. thats the 1st time my (non-login) name has been in a book. i've been approached to write or help on a book 3 times now (Web Services, Tablet PC, Compact Framework). these opportunities always come up right when i start a contract, and never when i'm off contract for long periods of times, so i have had to turn them down. maybe someday

wanted to get WS-SecureConversation working again but i cant get the dang sample to run. had to try some other stuff instead. being able to encrypt with a UsernameToken is one new feature for Release. played around with the samples and was able to get them to work until i switched the encryption to TripleDES. the client still seemed to work, but the server was complaining about the algorithm. saved off the client request and was able to write the CF code to decompose it. cant really write the CF piece to create it until i get the server side working. moved on to the CustomBinarySecurityToken sample next. its basically like WS-SecureConversation. you have to make a call to a token issuer, and then you can make calls to the web service itself. only had to make minor changes from the existing SecureConversation code to get it to work. hopefully that is a good sign that my existing SecureConversation code wont have to change much :)

the latest find is that WSE now signs the data after encrypting it on the server (tech preview used to sign and then encrypt). WSS specification supports both. the WSE 2.0 clients are still signing 1st and then encrypting, so this means the user will still 'see what they are signing'. i.e. you dont want to sign encrypted data, because you cant tell what you are signing. but i guess this legal matter doesnt make sense for a server? they cant really see ... yet. tried some stuff in code to see if it was customizable, but have not gotten anything to work yet. this last find closes the WS-Security loop for the CF bits. the code has been updated to handle the WSE 2.0 changes. it tested out doing a full cycle of signing and encrypting a request to the server, and then decrypting and verifying the signature of the response. now that i have WS-Security, should be able to get WS-SecureConversation working again.

get to the subway station. very few cars there so i pull out my PPC to make sure it wasnt a weekend. ok, just a slow friday. subway arrives and i actually get a seat. starts filling up at stops so i start watching the door for females to give up my seat to. believe it or not ... i treat women great. there is actually an open one next to me as well. next stop and 2 asian ladies get on, so i get up to let them sit. 2 guys that were already onboard and standing to the other side of me jump into the seats. adrenaline releases and i have to control myself from pummeling them. i'm new to public transportation, so i hope this is just a boston thing. so these 2 guys were low quality, but i also think womens lib is at work. the problem is i was going for chivaly, but the 2 other guys were going for womens lib. i'm pro womens lib too, but more pro chivalry, so when conflicts like this happen i get really pissed off.

msmobiles.com is reporting: REVOLUTION ARRIVED: Pocket C# - C# compiler running under Windows CE! hyping it up so that you can develop applications directly on your mobile device. ... that is a stupid idea. ... but there is something else. this would let us have a CF program generate code files on a device, assemble it, load that assembly into the executing program, and then run that IL. this is what happens on the desktop to make serialization (aka web services) faster. this doesnt happen on devices because they cant build assemblies ... until now

back to the problem about AES being the default crypto for WSE 2.0. bad news is that it does not exist for PPC 2003 or SP 2003. good news is that it will be in the next version of CE. more good news is that you can switch it back to TripleDES on the service using code or policy. bad news, if you use the code method, is that it seems to get ignored at times. this might be a web app lifetime issue ... but i've been experiencing odd behavior. have specific code on the client now to throw an exception when encountering AES and that has been hit more times then it should. anyways, that isnt good news at all because you are not likely to have control of the server. i thought WS-Policy might be able to let the client negotiate with the server that it can only do TripleDES, but it looks like i thought wrong. now i am thinking that WS-MetadataTransfer might be the spec i'll need. the problem then becomes that its not implemented ... nor published. to make this viable, i really need AES on the client. had ported the MSDN AES article code over to CF, but cant figure out how to make it compatible with the WSE2 AES implementation. not to mention any pure managed crypto is going to be dog slow on a Smartphone. if i find a native implementation to wrap, then it becomes an issue of getting it built for the different devices and emulators for testing. ugly

if you are lifting heavy weights, then sooner or later you are going to hurt yourself (not on purpose). first off, you have to heal, and there are some tips to help heal faster. ice is key. put an ice pack on the injured area periodically throughout the day, for about 15 minutes each time. ice will make your body pump blood to the injured area for heat, and your cells will use that enriched blood to heal whatever you hurt. elevation is also key. i'm not really sure why? something about blood circulation too. e.g. if you hurt your left shoulder, you might try to sleep on your right shoulder. when you start getting better, use heat and compression and work into it slowly. the main thing is do not immediately re-tweak your injury. you want to be careful, and eventually you will forget the injury and be good as new. second, there are tricks to maintain your health while injured. the main thing is that your body tries to stay symmetrical. so that if you injure your left bicep, then concentrate on your right bicep AND also your left arms tricep. i.e. you usually only injure one part of a push / pull muscle group; so if you hurt the muscle for pushing, then work on the muscle for pulling. this also works with chest / back and quadriceps / hamstrings. sadly, this concept doesnt really work for upper / lower body strength. anyways, that will help you maintrain while you nurse your self-inflicted wounds. the good news is that even if you are not able to maintain your strength, it is easier to get it back the 2nd time. serious, regaining lost strength takes a much shorter time period than getting stronger then you have ever been.

just cracked some more WSE 2.0 Release changes. Byron Kim and Krill Gavrylyuk (Thanks!) showed me how to default the Symmetric Encryption algorithm back to TripleDES (from AES128). this shows that the Xml-Encryption algorithm is at least partially working. Xml-Signature is working now for UsernameToken signing. the algorithm for generating a signing key from a password now only needs 16 bytes of data, instead of 24 for the Tech Preview. yes, i pulled that directly out of my ass. could we please get a readme.htm for interoperability changes? a simple note about that would have saved me 24 hours of BS. also, when i set the tracing to verbose, please tell me explicitly what part of the signature fails. tell me if it is the syntax, or a referenced element, or the signature itself. would be extremely helpful for interop dev. some new things too: added in the hooks for Filters, so the updated lib can call the CustomFilters sample. also cooked up the sample code to call the ContentBasedRouting sample. next order of business is to further text XmlEnc and XmlSig together, plus make the necessary changes for WS-SecureConversation.

a little help, please. the CF DIME bits seems to have been used out in-the-wild quite a bit. doing an update of those bits now, and am looking for bugs / enhancements that i can fix / push into the next release. please send me an email if you have found something. Thank you

for XmlEncryption, WSE 2.0 now uses AES128 by default (used to be TripleDes). the problem for CF is that AES does not exist in the CryptoApi for current mobile devices. need to figure out how to switch the service back to 3DES? for XmlSignature, instead of individually signing the Created and Expires elements, it now just signs their parent Timestamp element. also, UsernameTokens can now be used to sign and encrypt (used to just be able to sign). this is done by generating a key from the password. believe that the algorithm has changed in some form because this has stopped working from my Tech Preview bits? its a real pain to debug crypto code too. the most detailed messages you get are either BAD_DATA or 2 byte arrays that dont match. once i get past those 2 bugs, then the CF WSE bits will be back to their previous level of functionality.

err, umm ... why did anybody think of it at all. my spider has found some weird fetish sites: balloons, casts, others i wont mention. but it just found some new vids and carstuckgirls.com is the new champ in my book. the main page has some provacative clothing, but no nudity. not sure if it has any nudity at all. crazy stuff

the title is the name of a cardio class offered at my new gym. it involves a cardio workout while learning to do a strip tease. serious! tried to sign up for the class ... but they wouldn't let me :( also asked them if i could just watch ... nope :( not even if i brought singles? this gym also has a girls only workout room, which i run across from periodically. not that i care, but i'm assuming you would have women protesting if there was a guys only workout room?

Bill thinks he couldn't be less Well Rounded. i felt this same way when studying for the MCAD exams. there were certain questions that were going to be on the exam that i had never cared to learn, so i memorized them for the exam, and then purged them from my memory soon afterwards. e.g. i know web services like a champ, but ask me about Remoting and i'll laugh at you. never learned it on purpose. but that is scoped at the exam level, a bigger problem exists. this is exemplified by recruiters calling me up and saying we've got a project for .NET and C#. when i try to narrow it down by asking if it is Web Services or Compact Framework ... they have no clue. they dont know what TYPE of app it is. but certification is not geared towards a specialization ... the cliche 'jack of all trades, master of nothing'. it assumes that since i know C# then i should be equally qualified to develop desktop apps, web services, web forms, or mobile apps. what a joke. somebody fix this.

... is lifting as much weight as you can for 1 rep. for powerlifting meets, the lifts include: bench, deadlift, and squat. geared toward leg strength since 2 of the exercises are mostly lower body. you should be able to lift more with your legs than with your upper body, and peoples squat and deadlift maxes are usually similar. you should not max out often as it is hard on your body. it is easy to injure yourself during a max as well as it takes a long time to recover. you also have to train to prepare to max out. if you are working out consistently with 200 pounds for 10 reps, then you need to slowly increase weight and lower the reps. so next time do 220 lbs for 8 reps, then 240 for six, etc... this trains your body to get used to heavier weights, because the weight will feel different and your joints / secondary muscles will need to get stronger. once you get to sets with less than 5 reps, then you can think about maxing out. i dont max out anymore at all. too old. instead i have learned over the years how to accurately guesstimate what i would max. this kicks into play when doing sets of 5 reps or less. for every rep i do, then that means my max is +10 pounds more. e.g. if i bench 300 lbs. for 5 reps, then my max is probably 350 (300 + 5 * 10). for lower body it might be closer to +15 for each rep, depending mainly on how strong my midsection is. use this technique to judge my own strength goals. its also a good way to guess what other people max too. this system does NOT work for me on anything over 5 reps. that is because lactic acid starts coming into play. then it is more about endurance than actual strength, and it will vary from workout to workout depending on energy level.

boston offers a bunch of public transportation options. and it provides a handheld schedule planner to run on your pocket pc. what makes it high quality is that it is built with the Compact Framework! now if they just provided the subway schedule i need

still on track. upper body is starting to fill out. back is starting to feel snug in shirts and man-boobs (pecs) are getting some depth. have been concentrating on lower body for the time being. dont want to be one of those top-heavy gym guys people make fun of. leg strength is slowly coming back, but i dont feel a solid base yet. as leg strength returns my center of gravity will basically shift lower. dont really know how to describe it other than that. did neglect cardio and abs last week. need to get those worked back into the new gym routine. need to kick my bodyweight out of stasis too ... still hovering around 220. that means i could continue as-is and keep losing fat and gaining muscle. would be fine but my frame can hold more weight. so i'm going to slightly up the calories with more protein (at breakfast) to get in the 225+ range.

been updating the /spWSE bits for the WSE 2.0 release. have got WS-Addressing, WS-Attachments, and minimal WS-Security working again. first thing i did was change its security dependency to OpenNETCF.System.Security. then i started updating namespace strings to match the new OASIS spec. the Timestamp element is now enclosed within a Security SoapHeader element. also the ReplyTo element replaces the From element for WS-Adressing. Addressing was giving me this exception: The To header must match the Actor attribute on this receiver. The problem was that the sample web service has this decoration: [SoapActor("http://localhost/RouterService/StockService.asmx")]. this will not work for CF, since CF must use the machineName or ipAddress. it worked after updating that and the paths in the referralCache.config, which also was using localhost. will release the updated code once its ready. QUESTIONS: when are the microsoft.com web services going to start using WSE 2.0? when is MapPoint going to start using WSE at all?

a friend wrote me and said they didnt believe i had the proCreation fish symbol on the back of my car. so here is a pic as proof. there are actually 3 items: 1) proCreation fish 2) devil fish 3) .NET Framework bumper sticker! the proCreation fish is props to darwin (and humping). this is my 3rd one. it was ripped off parked at work and at a car maintenance shop; both in dallas. have a spare in the glove box. the devil fish is just funny. obviously dont believe in the devil. bought it the 2nd time the proCreation fish was ripped off. its been taped over about 5 times. have had the .NET FX sticker on for about 3 years. since its MS, its probably more evil than the other 2 combined :) no java or linux bigots have tampered with it

why does the Whidbey preview (May) break my project files from being opened in VS.NET 2003? the settings for both versions should be able to live peacefully together in the same file. not to mention it would ease my adoption.

most people dont know the difference. bodybuilders are trying to get big and lean. mostly asthetics. powerlifters are mainly trying to get strong, and getting bigger is just a byproduct. in general, bodybuilders will have larger muscles than powerlifters, but powerlifters will be stronger. you have to workout different for each. bodybuilders do lighter weights for more reps. they really concentrate on specific muscles, to the point of flexing a muscle while lifting. this is really hard to to, but a great way to make your muscle grow. they use strict form to isolate muscles and will workout the same muscle from different angles for max size and definition. on the other hand, powerlifters mainly target the large muscle groups with core lifts: squat, bench, deadlift, lat pulldowns. they lift heavier weights for less reps. they still control the weight, but can concentrate less on a specific muscle and more on moving the weight. other differences are that bodybuilders have to worry alot about their diet to get lean and powerlifters have to worry about not injuring themselves. powerlifters that cross over to bodybuilding usually do the best in competitions. powerlifting builds up the large muscle mass to wow the crowd, and then bodybuilding finishes it off for the judges.

an article on MSDN, Going to a Trade Show, is anti booth babe. in the 'tacky' section. booth babes can be smart too, and make perfect sense in our male dominated industry. there should be more of them

[why i dont believe in god] growing up, I had alot of confrontations with TX bible beaters trying to convert me. as an atheist, i didnt have the bible or faith to fall back on in arguments ... so had to come up with my own logic to not believe. its called thinking for yourself. tough trying to 'reason' with someone about 'faith' ... it just doesnt work. faith arguments always have a built-in escape door. instead, what worked best was 'what if' scenarios. have not run through this in a while, but these are the points i can remember: what if there is a god and what if there is a heaven. he (god) could be chaotic. so whatever you do during your life doesnt really matter. all that matters is the whim of your god. whatever you consider proof could just be said god messing with us. suppose he is rational. disregard a rational god creating primarily irrational beings. and he has an ego. that is why he created us in his likeness. so i live a good life, and he doesnt let me into heaven because i did not believe in him. in that case, i dont want that ego god to be my god. would rather have a god that wanted me to be good for the sake of being good, and not to get into his exclusive club. if he has no ego, and i live a good life, then he would let me into heaven and i would then believe in him. if there is no heaven. then what reward are you living a good life for? i think it is right to live a good life just for the sole purpose of living a good life. or are you being good to avoid the negative? if there is a hell. so the ego god wont let me into heaven and sends me to hell. then it depends on the devil. the devil could torture my soul eternally, probably for not believing in him (devil), just like the egotistical god. that would suck. or the devil would reward me for giving god lovers hell on earth. we would hang out, read blogs, and torture the god lovers that spoke the gospel but didnt act accordingly. if there is a purgatory. this is probably where i would go. dont believe in heaven or hell, so purgatory sounds right. my kind of place too. could hang out with all the other good atheists. likely to be a bunch of smart and interesting souls (or whatever) there. if there is no god / devil / heaven / hell / purgatory. then you just die and are worm food (what i believe in), forget the soul nonsense. so out of those, the only real negative is a rational egotistical god that would send me to hell, and that the devil would have to disregard that i was not god-happy on earth, and ignore that i was a good person, to torture me forever. if that really is the case ... then give me hell

taking the subway into work. it cut a full 15 minutes off my 1:15 minute commute. this is my 1st time to ever ride a subway, and here is what i've noticed: boston needs more parking. i start at an endpoint, but by the time i reach downtown, the subway is FULL, because you HAVE to commute. parking is so bad, i paid $28 to park all day over the holiday weekend ... and that was with validation. ridiculous. when exiting the city, i saw one subway get fully packed. i'm talking sardine style. i cannot read during the trip because i have to give up my seat to women. have yet to see any other males give up their seat to a female. asking around, a seat will be given up for the elderly and pregnant women. most of the guys pretend to be reading and wont look up to see the women having to stand. some that do stand have actually acquired the ability to read the newspaper while standing up and holding onto a pole for stability. and the seats are too narrow. when i have gotten a seat, i'm sandwiched between 2 strangers. i'm not very narrow myself, but with the pigging up of america, this is a bad scene. i've developed the ability to jab people in their leg with my pocket pc while it is still in my pocket. can do this subtley by directing a little pressure while pretending to rest my elbow on my leg. this will get them to relinquish some leg space so that the chesnut jewels can get some breathing room down there.

Chris starts A whole new category! the quote is classic. and i'm ordering the thong. err, umm ... not for myself. cough. i've got this car plaque on the back of my car: ProCreation. funny for multiple reasons. it's been ripped off or covered over a couple of times. now i keep a spare handy in the glove box. the best thing is seeing peoples reactions in the rearview mirror.

bodyweight exercises are great. this includes push ups, pull ups, dips, chin ups. if you are moving around your own bodyweight, your body will want to keep the muscle and get rid of the fat. it wants to keep the muscle so that you can continue to move yourself around. and it wants to get rid of the fat to make you lighter and easier to move around. note that these are hard exercises. you wont see too many people doing pull ups in the gym. especially bodybuilders, because they have weighted themselves down with muscle. i started out only being able to do 1 pull up. would go to failure each workout, and would periodically be able to do another rep. alot of gyms have machines that are weight-assisted, so they provide a counter-balance and make you lighter. use these machines until you can manage 1 unassisted rep, and then get rid of that crutch.

finished the book Fuzzy Logic (1993). got it for something like 3 bucks in Milwaukee. it had way too much history in it, although i did find it amusing (and a good way to knock some of the dust off after so much studying for certification exams). it did go over some of the tradeoffs between fuzzy logic, bayesian, and neural nets. that is what i'm starting to struggle with. there are a bunch of different AI techniques, so now i need to learn when to apply a specific methodology to a specific problem. need to stop reading and actually start coding this stuff too.

been catching up on blog reading. happy that i'm not the only one offending 9 out of 10 people: Finding Jesus at TechEd. they used to be real aggressive at the bible belt ... that is what turned me into a 'militant atheist'. dont have time to read ALL the comments (something like 300K of source!), but the 1st one is classic "Or worse, maybe this is just the real you". LOL! would the real Rory please stand up. hate to admit this, but i think Rory's blog has offended more people than mine. agree with alot of his stuff. dont directly agree with his reasoning for not believing, at least from what is in his post. not sure what all is in the comments? will have to post my own reasons for not believing later. actually, 'not believing' is a stupid way of saying 'being rational'

have had 2 good workouts after taking a week off from the TS visit. the weights feel a little heavier, but my energy level is back up. should be back to where i left off (or stronger) by next workouts. these workouts were just shocking the muscles to let them know that the vacation is over. had to find a new gym since relocating. initial observation shows that i'm in the top 3 for strength. the last gym had a bunch of beasts, i might have made top 10 in my current state. regardless, unless the alpha male hasn't shown himself, i should have this gym conquered before the month is out.

my 10 year high school reunion is tomorrow. never had any intention of going. dont want to know what my old buddies do for a living or that they have responsibilities now. dont want to see all the girls that wouldnt sleep with me. just so happens i was a bad ass back then, but i dont get the whole relive your glory days thing. i'm all about 'what have you done lately?'

having to do c++ at work right now (c# for the last 3.5 years, java before that). ouch! include paths, linking libraries, string conversions, etc... bump up the difficulty by having to support multiple devices. eVC isnt the greatest IDE, lack of consistent libraries, and targeting different CPUs. productivity has dropped something like 10x at the minimum. entirely humiliating experience. almost makes me want to spend my spare time re-learning c++ ... F that. waste of time. let the beatings continue

Big Bucks for Biometric Screening is basically what i was doing for my 1st job; as a summer co-op for Raytheon (previously E-Systems) out of high school. Raytheon's name even appears in the article. the 1st summer (right before college) gave me access to all the equipment necessary to make the best fake IDs you have ever seen :) it became a tradition for the senior co-op to pass that knowledge on to the intern their last day before heading off to school. my best alias ... john smythe.

took this pic about a week ago at Faneuil Hall (can i buy a consonant?). put myself in harms way to get that shot :) the sad thing is it was jacket weather yesterday. dont think i've ever worn a jacket in june? global warming my a55

the TS and i attacked boston over the long weekend. got to see a good chunk of the city. yesterday i relocated to a different hotel ... with internet access! so maybe i can get some work done this month? actually, last month i blogged my goals for what i wanted to accomplish, and all went well. had this imaginary feeling that my imaginary blog readers would hold me accountable :) going to repeat that pattern this month. brawn is still priority. need to start looking at gyms tonight. secondary goal is to get a new (updated) article put together. need to figure out which idea(s) to pursue. third goal is to knock out the MCSD requirements exam. go!

this is just over the 3 week mark of the brawn revival. last 2 workouts were great. did legs on monday to the point i was walking around like an old man yesterday. chest and back last night. managed 315 for 6 reps on decline bench and a set of 30 reps on dips right after. right shoulder is still a little sore, but not bad. TS shows up today for about a week, so she'll keep me out of the gym. it is good to take a week off every once in a while because your body will heal more and you will actually come back stronger and with more energy.

would love to be at this one: Going Mobile with .NET, XML and SOAP. somebody please take notes for me :)

it is out! this gives me cause to start updating the /spWSE bits. wonder if this means i'll start getting recruiter calls for my WSE experience?

Mr. Sells was spouting about The Next Big Leap in Programming. one thing that came out of it was Genetic Programming. certainly agree ... i'm just not that far yet. still trying to build up a basic AI skill set. but when i think about the near future ... AI keeps coming up. if you look at the Google career pages, they are all about AI and really efficient computing. and the parts that interest me in Longhorn will have some AI behind them as well. e.g. natural language searching. always try to keep a timeline of what is needed. security is needed now. i think web services are next, then mobile. after that ... AI. 3 down ... 1 to go. [grid computing is interesting to me as well ... but i dont see it being accessible] also have to think Chris for pointing out the MIT OpenCourseWare. outstanding! just so happens they have some AI courses for me.

get a number of email questions about articles i have written. always struggle with what to do about that, because they are typically repeat questions. tried to battle this by adding a comments link to the articles. then i could easily provide additional info there. problem is ... nobody clicks those links. should have made my comments render at the bottom of the pages. would prefer that people would just ask questions in newsgroups. just answer the question once, and then it is searchable for others. there is also a much larger audience to field the answer. what about the blog? i could answer questions on the blog. just doesnt seem as useful as the newsgroups. not well categorized, the posts have a short lifetime, and it would add some more noise. from the summit, saw that MS was working on making newsgroups a better experience. are they extending NNTP to do that, or are they just embedding their own metadata in the messages?

went to kinkos around noon to work on my website ... and it was blocked :)

You have attempted to access a site
that is prohibited or has been blocked
pursuant to Kinko's policy.
SmartFilter 

nope, not steriods. i'm talking about those jack lalange (sp?) juicers for fresh fruit and veggie drinks. used to do this hardcore, and i recommend it. i felt invincible when drinking fresh juice. first, i got stronger because i was giving my body more of what it needed. and if you think about it, the strongest land animal (elephant) is a vegetarian. second, i did not get sick. was traveling alot at this time, and people were dropping off left and right being sick, but i didnt even sniffle. thinking i could have withstood ebola. some tips: you cannot mess up fruit drinks. just pick a fruit and juice it, it will taste great. be careful mixing fruits because of the different acidities, so stick to 1 or 2 fruits at a time. veggies are less pleasant. the mix i liked was carrots, celery, tomato, and spinach. it looked like pond water but tasted better than pond water. while drinking it, i would feel good, it was like my body was releasing endorphins to thank me. that said, i did not like any other veggie mixes. some people do straight carrot juice. other would put an apple in to make it somewhat sweet. in general, i dont recommend mixing fruits with veggies ... ever. the downside about juicers is the time. you have to keep the fruit handy, juice it, and then clean the machine. too much trouble while traveling. some juicing mistakes i made: 1) tried to do an orange-strawberry-banana drink ... bananas dont juice well as all. 2) tried juicing broccoli ... did an entire head of broccoli, which amounted to about a shot glass of juice. THE WORST THING I HAVE EVER DRANK! speaking of shots, there was a juice place in colorado where they would juice a shot of wheat grass for you. got addicted to that stuff

hey hey! Sam linked the CF crypto stuff. linking back ... but i figure the chances of somebody being subscribed to my blog and not his is about 0%. for the longest time i thought his last name was pronounced like the merriam-webster word. jealous that my last name does not mean heathen or pagan :) think he was the 1st person to ever link one of my articles in a blog. note that he was spelling my last name incorrectly when i was pronouncing his wrong :) he just got an extremely cool press release. it was interesting that it pointed out his being an MVP. seems like the MVP branding might be starting to stick ...

passed with a 968. used the MsPress study guide. thought this one was adequate preparation because i've done very little COM+ and Remoting ... on purpose (sorry Ingo). looked at an Osborne study guide i had picked up from a $5 bookstore too ... it sucked. too many labs. actually, i dont like that the MsPress books have labs either. straight text would have been better for me. dont know if any of the study guides offer that? the programs with sample questions are valuable too. have never tried a transcender, but my assumption is they are alot of sample questions? would probably ace the exams with them ... but i dont want to spend the $$$. started studying for the 1st test just under 4 weeks ago, and now i've got MCAD. dont plan on immediately going for MCSD. couple reasons. 1st, i'm going out of my mind. cant read any more exam prep books or take anymore sample tests for a while. brain is going to mush. studying has helped, but i cant say i've learned anything useful. 2nd, i'm not interested in any of the tests i have left. the Requirements one is just a bunch of reading, and the electives are about 4 year old server products. Commerce Server ... dead. SQL Server ... get a DBA. Biztalk ... 2K version sucks. the study guide is even out of print. the 2004 flavor has me interested, so put together a cert exam. also, i dont think recruiters know the difference between MCAD and MCSD, but they do know the diff between 3 and 4 letter acronyms. thus, my theory is that recruiters think MCP and MVP are the same, and they probably get MCAD and MCSD confused too. so i'm golden now that i have 4 letters, because 4 letters is definitely better than 3 :) now we shall see about the ROI. cost me $250 cash so far (had a free pass to one). but a month of my free time was the main cost. would that time have been better spent doing something else? dont know. will let you know my opinion after some time passes. not sure what to do next. need to go back to building up an AI skill set. there are actually some DB related stuff i might kick around as well. will probably crank out an article before knocking out MCSD. in the short term i need to go celebrate and disconnect some of of these newly connected synapses.

got 2 emails from imagine cup groups the last couple days. my hotel internet connection does not allow me to send email, so i'm just going to post to the blog. from the /freeSpeech article, the BinaryCopy method is in the 'comments' link on that page. somehow that got left out of the codebase when i zipped it up over a year ago. also, that was the 1st port of DIME i did to CF. it only supports sending of 1 attachment. the latest port is in the /spWse article. it supports sending and receiving of multiple attachments, as well as closely follows the WSE implementation. and i repeat, the DIME code is a port from public code that MS made available. all i did was port it to CF and update it from an earlier DIME spec to work with WSE.

tired and feeling weak. been working out too hard lately. need to get more sleep and keep the calories up. what i dont want to happen is my body to start cannibalizing the muscle its been building up. luckily, its time for a rest day. aside, friend #1 read the blog from last week and asked how to join the casey sycophant cult. rule #1) there is no sycophant cult. rule #2) THERE IS NO SYCOPHANT CULT. rule #3) if this is your 1st day, then you must sycophant. further aside, sycophant sounds makes me think of a crazy elephant

if you didnt know, the Sys.Sec.Crypto namespace is now a part of OpenNETCF SDF 1.1. just got an email looking for an RSACryptoServiceProvider code sample. there are a number of code samples with the original source download of the article. there is a SmartPhone WinForms app that is the device test harness. it calls a WebService on the desktop to verify interop between crypto on the desktop and the device. you'll have to change the namespace from System.Security.Cryptograpy to OpenNETCF.Security.Cryptography. you will also want to read the article to find out some of the differences between crypto on the desktop and the device.

realized that i left out a caste. there are a lot of really smart people in boston too. BU, MIT, Harvard, ... oh my! not to mention the people i work with. an example, the security gaurd in my building was asking me if my mobile devices were bluetooth enabled. alot of tech here. have heard it called the silicon valley of the east. starting to believe it, although i've never actually been to silicon valley. aside, i sometimes refer to dallas as the 'silicone valley'. do you get it? switching back to the mean people. i'm guessing that is what it takes to start a revolution. saw a 'time squad' cartoon that seems even funnier now. basically it represented the revolutionaries as being dainty and drinking tea. then coffee was introduced and they got rude and revolted. this is where dunking donuts comes back into play. so there are alot of smart and big people here. could mean trouble for bNb. hope i dont meet somebody smarter AND bigger than me ... doh!

back to complaining about the streets. previously i was moaning about there being alot of streets because nothing runs straight for any great distance. at places they have compensated for this by naming connecting streets the same. so the same street has a bunch of 90 degree turns in it, and signs pointing you in the new direction of the steet. not physically connected, only logically through the use of minimal signs. i use the term 'logically' loosely. even worse, as my boss pointed out, and i can now confirm ... they have so many streets that alot of them in the metroplex are named the same. thus when you try to map it online, it comes up with about 4 streets in different locations for you to pick from. also, there are no on ramps. basically you have to go from a dead stop to highway speeds constantly. what little on ramp you could have had is always killed by a stop sign. and nobody changes into the fast lane to allow you to pull on and pick up speed ... more on that later. my conspiracy theory is that the makers of german sports cars paid off all the civil engineers. you see BMWs everywhere. just so happens that a good friend of mine lived in boston for 8 years and races BMWs. to top it off, out of the other 10 people i've met here, one of them races BMWs as well. coincidence ... i think not. i want a BMW now. not because of the streets, but because that article came out a couple weeks ago saying BMW owners had more sex. got to drive into downtown once. went over a bridge that had a tollbooth right in the middle. 1st time i've ever seen that. even got to make a short trip through the big dig. confused by that as well. my assumption was that the big dig was to increase traffic flow. but its not, its mainly asthetic. the tunnel handles just slightly more traffic than the elevated highways. serious, they are ripping down what they used to use, to make the city look better. this goes against most of what i believe in. what they should have done is built more parking so that i wouldnt have to take the train into downtown. although i really want to put my ear on the train track and see if i can tell how far away the train is or which direction it is coming from. tonto style. rule #1 of the train ... do not talk to anybody. everybody is reading. i just pretend to read to fit in. also adapting my speech to blend. this involves not pronouncing the letter 'r'. e.g. where i sould say 'car', they say 'ca' with a soft 'a'. had to go to speech therapy when i was a kid because i could not say r's properly; should have just moved here instead. other jargon: sneakers instead of shoes. soda (or tonic?) instead of coke. and dunkin donuts is called dunks by the cool people. i'm trying to work in a little sexual innuendo by calling them double D's. just a matter of time before it catches on. serious, there are coffee shops everywhere. been taking my certification exams by where one of the universities is. tons of coffee shops and they are packed. everytime i have been there it has been hippy central. if somebody was able to find grass (which is hard to do) then they are laying in it. consider myself open minded, but these people are crazy. little bit of a deviant, but in this place i basically have to conform to be different. did i mention the democratic convention is going to be here? that explains alot. hate the party system, but its much different than the republic of texas. i was not expecting the ultra liberals. from the stereotype i was expecting a bunch of smart asses. dont worry, i havent been let down. have the capability to be mean when necessary, but i rarely have to use it. been having to exercise it more often here. one person termed them a bunch of hard asses. agreed. related to this, everybody at my gym is huge. more weightlifters than any other city i've been to. theory is that they are getting big so they can be mean more often. size-wise i'm about average at this place (currently). alot of them are jacked up and heavily tattoed. strength-wise i'm closer to the top but still have some work to do. what else? ... i'm confused by the fire hydrants. the base is painted black, and they have either a red or yellow top. wonder if the color of the top has some significance, e.g. water pressure? also, why isnt the whole thing painted bright red or yellow? theory is that they want them harder to see so that the old buildings will burn down and then they can build newer ones. the new buildings should already have sprinkler systems in them, so the fire hydrants are less important.

1) my boss explained the brown eggs to me. only a certain breed of chicken (something reds?) can make it up here, and they lay brown eggs. should be able to find white eggs at a large grocery store, but they will have been imported. this disproves my theory about the continuance of the civil war between chicken farmers. 2) the electrical outlets being upside down in my apartment is just within this hotel. have not seen this mistake repeated elsewhere. 3) the flashing green street lights have something to do with pedestrian crosswalks. i need to walk around that area and take some pictures, so i'll investigate further at that time. hopefully, no tourists will be driving at that exact time and run me over while distracted by the flashing green lights.

my boston contract happens to be at Sam Gentile's Big Announcement. Phil Stanhope got me here. as bostonians say: "wicked cool"

it has been 2 weeks since making brawn a priority over brains. that isnt very long, but there are definite changes. the mirror test shows that my body weight is being distributed differently. starting to carry a little more muscle and a little less fat. the mirror test is priority over what you weigh. only girls and jockeys care about their body weight. that said, i weighed in at 220. that was my college playing weight and seems to be what i hover around as an adult. for my height that puts me right at the obese category for BMI (body mass index). BMI basically makes no sense for weight lifters. happy with that weight because it means i'm not losing. actually i'm losing fat but adding muscle (you cant turn fat into muscle). so i'm getting the right amount of calories, exercise, and rest. couple more weeks and the muscle will hit critical mass and start burning off the fat more rapidly. finally, the belt test (counting belt loops) shows that my waste is a little smaller. the body weight workout i just did showed promise too. was able to do a couple more reps with each exercise. squat workouts show that my max is probably low 400s and bench max would be 350. the only negative is that my right front shoulder is a little tender. need to be careful not to injure it. the TS is coming to visit in a week, so that will be some good recovery time when she is here. can tear myself down for this last week, heal while she is here, then be recharged to resume.

real tired from jumping 3 times zones, plus having a redeye flight back. the last i heard was that a light workout with weights will help you recover faster. heading to the gym now to do some bodyweight stuff (push ups, pull ups, dips) to knock off some of this lag. if i were to workout real hard then my body would just get shut down and get sick as a protective measure. aside, no clue how pro athletes can travel so much and stay in shape?

as soon as the plane took off to seattle, boston started with the gay marriages. have not seen any pics of lesbians getting married, so all is still good. flew alaska airlines for the 1st time. the girl in front of me was making some noise about not getting the vegetarian meal she had ordered. she ended up not eating out of protest. assuming it was worse for her body to not eat than if she would have ate (at least part of) the chicken meal? then i got my meal and it had a little prayer card with it (psalm 107). next time i'm going to order the secular meal ... or not eat. get to MS the next day. the HR person asked what i had done to prepare? err, umm ... brushed my teeth. nobody told me i was supposed to study! get shuttled over to another building to interview; Don Box walked out the door of the building as i was going in. see title above for the extent of our conversation. on the off chance Don reads my blog, and on the off chance he remembers a guy with a shaved head calling him mr. box ... then that was me (TS: sorry about the shaved head). kind of cool. while waiting, it stood out that EVERYBODY was either carrying a laptop or a backpack (unknown if the backpacks contained laptops). so i interview and the end result is they dont make an offer. whew! if they would have made an offer, then that would have been a BIG decision for me to make. although i am feeling a little bitter about being rejected, especially since they wont tell me what their logic was. would have made great blog material. maybe i should have studied :) the funny thing is that one of the interviewers told me he had read my previous post about MS interviews ... and then he asked me a riddle! classic. the long flight did provide alot of reading time. finished the cert study guide i took with me, so should be able to knock that test out this weekend. negative was that this trip hosed up my workout schedule. need to recover from the timezone changes and then dive back in.

heading out to MS for an interview tomorrow. will let you know what they say about my blog :) if they make an offer and if i hire on, of course i would clean it up. the funny thing is that some people have acted impressed when i tell them i'm going to interview there in general. i dont understand this. is it company worship? have heard that they are selective ... but 50K+ is alot of people to have been selected. other rumors are they dont pay as well. maybe this has changed with the fall out of stock, not to mention they have picked up alot of big named people. need more info on the redmond area itself. the weather doesnt look great, but manageable. sort of eternal fall. a major positive is technology. some have called it a geek paradise. that depends what you are working on, and how tightly focused you are on that problem, but there is alot of cool stuff happening there. they are bringing me up for technology X. they had a couple positions in tech X that they showed me, but only 1 was compelling. although i have tech Y too, no positions in those were offered up. i'm assuming that represents the problems MS has getting across product lines? also, the people. would be extremely cool to get to meet and learn from the people there. now to the interview rumors. riddles do not help you understand how people think. all they do is make the person trying to solve the riddle feel like an idiot. you want to make the candidate look bad so that they will be humbled and want the company to hire them just to improve their riddle solving abilities. i've heard of people memorizing all the riddles that they could find on the internet to prepare ... stupid. my riddle strategy is to think way outside of the box. they are like: "how do you tell which light bulb was on?". me: "do i have super powers?". them: "no". me: "not even a genetic mutation?". the only person i've seen that needed advanced riddle solving skills was batman. he would constantly have to solve them to catch the riddler. also heard about having to whiteboard code for data structures. another skill i've never had to execute at a job. employers prefer that i actually type the code in and use the System.Collections namespace :) [until Generics] MS research is probably already working on whiteboards with intellisense. tech interviews have changed for me since starting the website. if they want to see how i think, then they can just read an article. typically document the mistakes i made and what i had to teach myself to get something to work. i really wish other people would write like this too. most articles just write the steps to get it to work, and omit what went wrong. i think the wrong paths that you took are just as important to understand so that others will not follow. it also provides a more complete understanding. else they might follow and turn what you thought was a wrong path into an alternate route. then there is code. if you want to see some, i have smack loads available for download. if those niceties are out of the way, then they can spend all the time trying to figure out if i know how to be professional :) anyways, i wont be able to tell you what actually happens, and the above are just my thoughts about the common rumors

Jason Haley invited me out to a brewery tour in boston. odd that i never went to a brew tour in milwaukee? Jason was a TA when i presented .NET CF to harvard last year. finishing up his MCSD right now, so i bugged him with questions about that. got to see some familiar faces ... and drink beer. picked their brains about the nuances of boston, which i'll be posting later.

nope, not me. those are the people against my blog. you only ever hear about the devils advocates though ... why is that? there is too much to comment on in an organized manner, so i'm just going to knock them out as i go. 1st off, i appreciated that the recruiter wrote me back and was honest. i'm honest, so i respect honesty. did i feel discriminated against ... hell no. i know exactly what i'm doing with my blog. have already been turned down for jobs because of it, and i expect to be turned down for many more. that email just happened to make it into the blog. am i upset about not getting the job ... not at all. do i think that i could have helped that company ... most definitely. do i think they are closed minded for not being able to get past my blog ... yes. but that is my opinion. they have their own opinion, and they expressed it. so what. would i ever sue ... hell no. i hate lawyers and i hate people that sue. if you are a lawyer reading this, i hate you. now on to the reality bit. the reality is i am not hurting for work. all of my needs are satisfied. since my needs are met, i can pursue ideals and wants, in that order. ideals because i have to live with myself. wants ... i dont have many. the world is easy and my life is quite simple. did you figure out i was a philosophy minor? people can assume whatever they want. so companies choose not to use me ... there are lots of companies. and i will help the companies that do choose to use me put the others out of business. boor: no, i am not rude. am i insensitive, you better believe it. i am quite hard on myself. maybe being offensive on my blog is a way to make things more challenging? being 28 ... my age is irrelevant. yes, i could make 2 web sites: 1 personal, 1 professional. but i dont care to. i like having 1 site; especially when entertaining stuff like this happens. could cut out the racy stuff at any time and this would rapidly be forgotten. but what have i dont that is really bad ... absolutely nothing, my site is so tame. people just like to feel better than others. professional references are too easy. i've never been kicked off a contract, or even been close. the case has been that they want me to stay, but i get bored out of my mind and have to move on to something else. or the weather gets cold and i have to migrate south. yes, i want to see more people posting pictures of their girls butt. i'm a guy ... cant help it. offensive topics to who? there are people that blog pro-christianity topics ... dont you think i find those offensive as an atheist. or are the xtians the only ones that get to be offended? honestly, i dont find them offensive. dont care. this is a shock blog. if you cannot handle shocking things, then dont read it. recall telling people to unsubscribe many times. and i do point employers to it ... on purpose ... i've already been through all this crap. i only want to work for the people that can handle my blog. they are not filtering me ... i am filtering them. a couple years ago i had the adult web sites that i had worked on my resume. ended up working with the greatest bunch of guys because of that. they were smart, hard working, professional ... but could have fun too. chances are this will continue. you followed your path, i will blaze my own. this is my life ... i am in complete control

funny stuff. bill got a negative comment on his blog about my blog. about time the silent majority spoke up. bill has already ran through it. need to arrange for bill to be my official spokesperson :) he does a much better job of representing me than i could myself. serious. got to go to work now ... but this is high quality. would love to hear the MS recruiters opinion

[cant sleep] my car radio makes the car. all i care about the car is that it is reliable and gets me from point A to B ... and that it can carry speakers. loud bass is my sort of 'music that soothes the savage beast' bit. current setup is an mp3 cd player with 2 amps. 1 of those amps is dedicated for subs along with a cap for extra power. the best thing about my sub amp is that it has a wired dial on it. mounted the dial next to where i rest my hand while driving for easy access. with this setup, i can max out my amp settings in the trunk, and then adjust the bass song by song. this is a must have. it allows you to tweak each song for the right amount of bass. end up using this bass dial more than the actual controls on the radio itself. it also allows me to cut off the bass if a cop drives by. now for ideas i want to see implemented (a friend of mine from pittsburgh helped flesh out these ideas) ... IBS. instant bass switch. aftermarket car radios should have a button to send a pure bass note to the speakers. the scenario is ... i'm at a stop light playing some mellow music. then some punk kid rolls up next to me trying to bass. instead of trying to switch songs real quick, i could press the IBS and let them know what i'm carrying. this would also replace my horn since my car horn is really weak. could just bass at people when they cut me off in traffic. next up is the Defcon system. my car is as loud as it needs to get. if i want to talk to somebody in my passenger seat, i can only turn the volume up to 1 or 2 ... it goes up to 80. an average bass session has the volume up to 8. 15 starts to get uncomfortable. the bass is usually dialed down to around 25 to 50%. so my car has alot more bass left in it. the defcon system would show this. basically it would record the level of your car at its loudest into memory. then as you drove around it would show your defcon system on the rear window or something. if you had your car at its loudest, then it would show defcon 1. mine would typically show a reading of 4 or 5. this way bass=mobiles could quickly size each other up. you could easily tell if somebody was bassing at their full potential ... or if they were just getting warmed up.

feed the machine. if you are working out hard, then you have to eat the right things to support that. this knocks atkins out the door, because it wont let you get a balanced meal. not to mention mr atkins was fat. carbs are great energy, energy that you can use to workout and gain muscle. what matters is when you eat those carbs. e.g. carbs right before bed is stupid because that energy will not get burned and it will become fat. eat your carbs at breakfast or lunch. i reserve protein, green veggies, and dairy for evening. fruits and veggies are great. somebody said 'have you ever heard of somebody getting fat off of them?'. me neither. to get enough, i have to eat them all day. fruits mostly in the morning because of the sugar, and veggies towards the evening. protein is required all day just so i can get enough. its somewhat impossible for a weight lifter to get too much protein. cut out the coffee and soda as another thing, caffeine makes the bones weak. also, for those calories, wouldnt you rather be eating something instead? water is life, and muscle needs alot of it. the no calorie thing is just a great bonus. if you are drinking alot of water too, then your body will stop storing as much, so you will get leaner. that brings up something about creatine. some creatine powders will end up putting a layer of water weight on you. used to add about 5+ pounds of water to me, that would go away when i stopped taking it. dont get that side effect with creatine clear. so i am wary of carbs and sugars, but i still eat them; but mostly in the beginning of the day. the body needs protein to build itself back up at night, so that is what it gets. as posted earlier, the lots of small meals throughout the day thing seems to really work. you will want smaller portions and your energy level will remain constant. the only hard part is sticking to it within our 3 meal culture. which brings up, you have to eat breakfast. you will need that energy later on in the day, and it gets your metabolism going to burn calories. so i dont diet in the traditional sense. if you starve yourself you are going to lose fat and muscle. so then you are a skinny wimp. the new methodolgy requires that you gain muscle to help you burn off that fat. this is how bodybuilders (lots of muscle) are able to get so lean. to support that, you have to eat. you just have to be careful what you eat and when you eat it. personally, i dont want to lose weight. i want to stay the same, and just trade fat for muscle. all that being said ... cheat day! once a week, you should eat whatever you want. i do this because it is fun. but the logic is that if you keep everything low fat, then your body will start hoarding the small amount of fat that you do supply it. if you shock it once a week, then it is supposed to remind your body to keep up the work. also, it is supposed to be a reward for your will power through the rest of the week. if you really want that donut one day, skip it and have one on cheat day instead.

passed with an 830 something. not happy with the score, but it's still a W in the win column. the score sheet shows that i totally tanked the 'creating user services' section and aced the rest. not happy with my study guide for this one. used the MsPress book like i did for the web app test. the web study guide was certainly thicker, and i picked up a couple 'gimme' answers from it. while reading this book, all i got was a false sense of security. was like 'i know all this crap'; so i made a 3 day turnaround to take this one. probably would have made the same score without studying at all. will try a study guide from a different publisher for