The TopCoder algorithm competitions were entertaining, and I learned a lot. The marathon matches were grueling, and for the first while interesting, and I learned a fair amount.
The dev/design competitions were, for me, quite painful. It was an incredible chore to force my self into using their process, and as I was banking on it for some income, the discovery that solutions were rejected because I was missing absurd amounts of documentation/unit tests meant I don't see myself going back to it any time soon.
The design & dev competitions began to overwhelm the community as well. I haven't played there for a while.
Agreed. I'd call myself a fairly capable programmer, althouth not an expert in their main technologies (.NET and Java) and from the few times I attempted to compete in component development, the amount of work involved just seemed a bit much. (especially the amount of unit tests) I'm not sure where the article's author got the "1-2 components per week, part time" figure from. It seemed to take me almost a week full time to complete one $500 component. (which, inevitably, didn't win)
I'm not complaining I didn't win: the judges clearly thought the other competitors did a better job, and they were probably right. But it's anything but easy money.
TopCoder seems to be overrun with developers from developing (largely Asian, as far as I can tell) countries where salaries (and cost of living) is much, MUCH lower, which is also keeping the prize money low. $500 (before tax) for a week's worth of work is just not attractive to me, I can make (and have made) 3-5x that doing normal consulting. (where the probability of being paid in the first place is a lot higher)
There is, of course, nothing actually wrong with this. This may well be the best way to outsource development to cheap labour: if it's not good enough, don't pay them anything!
There's probably a certain amount of practice to this, over time you figure out what the judges want, and what not to spend time on, but to make a competitive amount of money, you'd still have to have a very high success rate. The article is somewhat misleading in this respect.
That said, if you're in it for the fun, it's great!
How is this immoral? They tell you how it works upfront, they pay you one-off for winning the competition, and they pay you royalties when companies buy the software. It would be immoral to judge the best entry as being not up to scratch, not pay out, and still sell it on, but that's not what happens. (the contest is re-run in that case)
Besides, it's what you call a (sustainable) business model.
The dev/design competitions were, for me, quite painful. It was an incredible chore to force my self into using their process, and as I was banking on it for some income, the discovery that solutions were rejected because I was missing absurd amounts of documentation/unit tests meant I don't see myself going back to it any time soon.
The design & dev competitions began to overwhelm the community as well. I haven't played there for a while.