rss
 
comment(s)

archives
J|F|M|A|M|J|J|A|S|O|N|D
(20##) 10 9 8 7 6 5 4 3 2 1 0 <
 
DesktopWeb FormText   coming up with (side) project ideasThu, 01 Sep 2005 02:37:19 GMT # 

first, i pick the technologies to learn or reinforce. this is the easy part. its usually what new technology has come out or what would i ultimately like to work on full time. it also involves avoiding the technologies i do not want to work with. this might start with reading a book about the technology or reading the API to see what it can do. right now i'm leaning towards reinforcing MCE and using DirectX for UI. could stand to add in some tech i've never used before ... to up the fun. more AI would be great too. second step, find a problem. usually check my ideas.txt file first to see if i've already come up with a cool idea in the past that could fit the bill. (serious, i have an ideas.txt file that i've been compiling for years. all sorts of craziness in that.) if thats dry, then i look at the newsgroups / forums to see what problems other people are having (usually the consumer groups and not the developers). still no luck, then i start browsing through books and reading the APIs to see if i can use it in some way it was never intended to be used. if that fails, then i take an overdose of ginko biloba and have extended brainstorm / surf sessions with the tablet. this step also involves doing a bunch of searches and making sure it hasn't already been done before. the current idea was spawned by a news article. its another persons idea that has not been implemented yet (as far as i know). haven't come up with a legit way to work in AI either. third, tweak the difficulty. if its too easy, think about ways to make it harder. e.g. using a beta product, writing it on a mobile device, or mixing technologies. if its too difficult, maybe only tackle the high risk stuff and dont follow through with the grunt code. the more side projects i do, the better i get at knowing if i can actually implement the solution. early on, alot of my side projects would fail; but now its becoming pretty rare. hopefully this doesnt mean that i'm taking less risks! current idea is probably too easy. fourth, tweak the timeline so that it can be done within a month. just so i dont get stuck in a rut or get bored with it. also so that if something super cool comes up in that time, then i can easily ignore it for just a little while and check it out when the current project is done. current idea could be done in 2 to 3 weeks, and almost all of that is grunt code. fifth, judge my commitment. if the problem doesn't consume me so that i have to solve it ... then back to the drawing board. current idea definitely fails in this regard. think that i'm actually going to start to code a piece of it though and see if anything interesting comes up ... sometimes the idea doesn't fully present itself until you set things in motion ...