Crowd sourcing Mesa development experiment

[chrisb]

I can't even explain how shitty of an idea this "competition" is. Take the people with full time jobs first: "Hm, should I give up a months worth of work, for the possibility of receiving a price afterwards, while running the risk of not being able to pay rent and feed my family?". Which leaves out only students as potential partakers. Then, the first one to implement a sloppy version that is badly documented, but still somehow works, gets the price. Practitioners would try to take as many short cuts as possible. That surely can't be considered "quality open source development".

What you describe is the same problem as "paid per job" contractors. There has to be some quality metrics and sometimes this approach is viewed as not suitable for all tasks, which is where pay per hour comes in. But there are obviously some tasks where it does work, like the prizes I previously mentioned. The challenge would be to find some similar tasks in the open source world, where the main problem is not something nebulous like "code quality" but, say, reverse engineering a protocol, and providing a basic implementation. Decss would be a good example, all people needed was the algorithm and a working implementation, nobody cared about the "code quality". Code quality is a difficult metric to measure, and many full time and paid per hour programmers write crappy code, there's no guarantee regardless of how you choose to fund it.

[archibald]

I can't even explain how shitty of an idea this "competition" is. Take the people with full time jobs first: "Hm, should I give up a months worth of work, for the possibility of receiving a price afterwards, while running the risk of not being able to pay rent and feed my family?". Which leaves out only students as potential partakers. Then, the first one to implement a sloppy version that is badly documented, but still somehow works, gets the price. Practitioners would try to take as many short cuts as possible. That surely can't be considered "quality open source development".

You might get a rushed, crappy and undocumented implementation, but at least you'd have LOTS of rushed, crappy and undocumented implementations! ;-)

[Ancurio]

A competition wouldn't just be bad for the losers, it'd be bad for the whole of mesa: by definition you'd have 10 people working on 1 feature. Wouldn't it be better (less duplication of effort, faster support for new OpenGL versions) to have them working on different extensions?

If it's a competition then anybody who can work on it full time (or take some time off work) would be at a significant advantage. I don't think that a competition will attract more people to it.

I can't even explain how shitty of an idea this "competition" is. Take the people with full time jobs first: "Hm, should I give up a months worth of work, for the possibility of receiving a price afterwards, while running the risk of not being able to pay rent and feed my family?". Which leaves out only students as potential partakers. Then, the first one to implement a sloppy version that is badly documented, but still somehow works, gets the price. Practitioners would try to take as many short cuts as possible. That surely can't be considered "quality open source development".

[chrisb]

While I can't stop anyone starting up a new project the problem I see with a competition (aside from the duplication of effort) is that you need a non profit org to sign on to the idea otherwise the prize money would likely get taxed twice, once when the organiser collected the money and again when it was paid to the dev. Just something to think about it you are going to follow up on the idea.

Yes, afaik it would be pretty trivial to set up a non-profit here. Whether anyone would trust it enough to send money is an open question. The receiver of the money would obviously have to pay any local income taxes that apply, but they are free to take any possible steps to minimize that.

[chrisb]

A competition wouldn't just be bad for the losers, it'd be bad for the whole of mesa: by definition you'd have 10 people working on 1 feature. Wouldn't it be better (less duplication of effort, faster support for new OpenGL versions) to have them working on different extensions?

If it's a competition then anybody who can work on it full time (or take some time off work) would be at a significant advantage. I don't think that a competition will attract more people to it.

Collaboration is good, but there is something about the prospect of winning a cash prize competition that motivates people to do it. Consider the Ansari X Prize or Netflix Prize or Orteig Prize.

[tarceri]

Well it worked for the x prize... I would be willing to chip in for a competition, but choose some other goal, I don't want to take over this guys project.

While I can't stop anyone starting up a new project the problem I see with a competition (aside from the duplication of effort) is that you need a non profit org to sign on to the idea otherwise the prize money would likely get taxed twice, once when the organiser collected the money and again when it was paid to the dev. Just something to think about it you are going to follow up on the idea.

[archibald]

A competition wouldn't just be bad for the losers, it'd be bad for the whole of mesa: by definition you'd have 10 people working on 1 feature. Wouldn't it be better (less duplication of effort, faster support for new OpenGL versions) to have them working on different extensions?

If it's a competition then anybody who can work on it full time (or take some time off work) would be at a significant advantage. I don't think that a competition will attract more people to it.

[chrisb]

I like the idea of competition for money. It'll suck for the losers, but that is the whole incentive to be the winner.

Well it worked for the x prize... I would be willing to chip in for a competition, but choose some other goal, I don't want to take over this guys project.

Edit: OTOH if he is good, then he should win anyway, right?

[duby229]

Perhaps we should have a little competition - first to implement the extension gets the money

I like the idea of competition for money. It'll suck for the losers, but that is the whole incentive to be the winner.

[deri]

I meant something like that I would like to know how the process goes on, when I am funding something and I would like to fund wine project too if I found someone who would support it. You got already 464$ on indiegogo. That's not bad.