Sponsor An Open Source Driver Dev For a $1 a Day???

[ryao]

I talked about a similar matter with Martin Peres of Nouveau recently with there being a certain company that's looking to help drive Linux and was wondering if hiring open-source GPU driver developers would help improve the situation, etc. According to Martin, however, it really wouldn't immediately improve Nouveau. Most of the Nouveau driver developers are European university students who aren't financially-strained and/or they're already employed elsewhere. It's not that there's currently a line of qualified developers waiting to contribute to Nouveau but can't afford to do so.

The solution is to hire new people, let them train themselves and then have them get started. People do not become qualified unless they learn. That is how the Nouveau developers got started.

[blackiwid]

OMG First I thought great idea, after reading this that it would mainly come to the nouvou driver guys, I hate that idea, why would we sponsor guys to make a good driver for a company that blocks such stuff so good it can and not release any specs or something, then we help that company to get deals like that in china posted here somewhere, so we work very hard or pay money so that the company that sucks dont get punished for sucking on this area?

Thats retarded, than I would advise to amd, stopp making your driver, you will get that for free if you just do nothing. We live in the capialism, if a company makes wrong desitions get not punished for that the last thing thats kind of works goes away too. Becuase normaly it should give a law that drivers (maybe even all stuff but shure drivers) should be made for free only, if you buy a hardware you pay for that... but thats not working in our finance-market-dictatorship so at least the market-mechanism should not be reverted by such fundings...

I am not really happen that the nouvou driver even exists but maybe its ok that you get a picture and some 2d accel if you use that, but more than that... you should not make more effort to bring it near the radeon driver quality... than it would be absurd to release any specs in the future for any company if it makes no difference...

[entropy]

OMG First I thought great idea, after reading this that it would mainly come to the nouvou driver guys, I hate that idea, why would we sponsor guys to make a good driver for a company that blocks such stuff so good it can and not release any specs or something, then we help that company to get deals like that in china posted here somewhere, so we work very hard or pay money so that the company that sucks dont get punished for sucking on this area?

Thats retarded, than I would advise to amd, stopp making your driver, you will get that for free if you just do nothing. We live in the capialism, if a company makes wrong desitions get not punished for that the last thing thats kind of works goes away too. Becuase normaly it should give a law that drivers (maybe even all stuff but shure drivers) should be made for free only, if you buy a hardware you pay for that... but thats not working in our finance-market-dictatorship so at least the market-mechanism should not be reverted by such fundings...

I am not really happen that the nouvou driver even exists but maybe its ok that you get a picture and some 2d accel if you use that, but more than that... you should not make more effort to bring it near the radeon driver quality... than it would be absurd to release any specs in the future for any company if it makes no difference...

As I see it, this sponsoring is project related instead dumping money into a single pool.
At least I hope so. That way anyone who likes to see Nouveau improvements is free to sponsor it.
The same goes for AMD related projects. Where's the problem?

[l0b0]

The key would be targeting a small set of cards, possibly just a single series from one or two vendors. Maybe a system similar to the Humble Indie Bundle could work:

1. Select your own payment, ¤0.01 upwards.
2. Choose which vendors/series/cards the developers should focus on, with an additional option of "Any" to let the rest of the community choose.
3. Choose which stats the developers should focus on, for example color fidelity, fast antialiasing, or simply FPS.
4. Give extra incentives to those who pay more, either with stages (like Kickstarter) or using the average (like HIB).

If the developers were able to boast something like 60 FPS in all the initial Steam for Linux releases on default settings on even a single card, that would be an enormous achievement, and it could be the wedge to finally level with Windows as a gaming platform.

[Deavir]

If AMD wanted to profit off of NVIDIA's recent PR issue, a kickstarter project where they match funds to hire programers to work on the open stack for 2 years could be started. The kickstarter side of the programmers time would be community focused by vote, AMD's paid side business as usual. Gifts would be easy from AMD/ATI stickers, videocards, to full systems on reward tiers.

The matching funds would be so people wouldn't complain as much that they were getting developers for free and not doing their part. Keep it in house so developers have the best access to hardware and intellectual resources. This would also take care of people worrying their money would go to hardware or features they don't want to support.

[entropy]

The key would be targeting a small set of cards, possibly just a single series from one or two vendors. Maybe a system similar to the Humble Indie Bundle could work:

1. Select your own payment, ¤0.01 upwards.
2. Choose which vendors/series/cards the developers should focus on, with an additional option of "Any" to let the rest of the community choose.
3. Choose which stats the developers should focus on, for example color fidelity, fast antialiasing, or simply FPS.
4. Give extra incentives to those who pay more, either with stages (like Kickstarter) or using the average (like HIB).

If the developers were able to boast something like 60 FPS in all the initial Steam for Linux releases on default settings on even a single card, that would be an enormous achievement, and it could be the wedge to finally level with Windows as a gaming platform.

Sounds good to me, although too fine-grained options may be a showstopper
in case not that many people are willing to give money.

Maybe it would be even make sense to approach some bigger companies and
ask them if they would also participate in a certain way like
"Doubling any sponsored dollar."

For instance, I'm thinking of companies like Canonical and VALVE.
Both companies may profit directly but also positive press coverage shouldn't be underestimated IMHO.

[l0b0]

I hope people agree that this is not a one-off project which is ever going to be "finished." To make the code future-proof, and to be able to offload smaller jobs to other developers, it is critical that the code is modular, testable and readable. The GSoC/EVoC method of hiring students works for short time, small budget projects, but to make this a long-term win there has to be top-notch developers in the core team.

The location is another important aspect - It's orders of magnitude more convenient, accurate and productive to just show someone your monitor than to turn off the lights, take a good quality photo (or even video), and then write a concise, understandable summary of the situation to someone who might not get around to even checking it out for hours. Never mind having to negotiate what needs to be done for the others to reproduce the issue. Therefore the core team members should all be willing to relocate (location to be negotiated) to work there for an indeterminate amount of time (essentially as long as the funding lasts/is renewed).

[dyrvere]

Feedback:
Anyway that’s the gist of my idea. I’m looking for feedback so that I can document more ideas around the subject and possibly put together a more complete proposal.

This I'd fraggin' do in a heartbeat as long as I can be sure it's for general Drivers as well, and not just for Graphics, I'll need a Bigfoot (Qualcomm Atheros) Killer E2100 (powered by Linux on the Network Processing Unit) driver to get at least fundamental network connection going (My motherboard came with one built in and I have to use a secondary Network Card) to start with if needed be. And not only 1 dollar a month, I'll go as far as 25 USD if possible to fund driver efforts on Linux! I could even purchase a card if I knew about someone willing to do the deed since I myself don't have a clue how it all works.
I'd also suggest to use some sort of Donation page where you can set the amount and use sliders to direct your fee towards certain types of categories like Network, Graphics, Sound, Input or what else could be. Something like Humble Bundle does it with Developers/Charity/Bundle tip.
Oh and for Pete's sake, the only reason I haven't been able to fund most Kickstarters is because they do not support PayPal. I do not have a credit card and I never intend to get one, I also dislike having to create virtual ones or borrow from others.

[Micket]

Which system for micropayments do you guys suggest?
Every system I've seen especially punishes small payments, and then a hefty percentage on top of that. Very discouraging.

[dogsleg]

Yes, the idea is in the air. I'm in. Many others here suggested close ideas. So, it is in a very need, almost necessary now.

Last time I suggested that:

I think that it will be very useful. Also I think it should be like an open auction. Users could suggest ideas to developers and developers could offer a price and a project-time (time to implement the suggested feature); then users could choose a developer with an appropriate experience, price and project-time and send him (her, them) money.

It is somehow an alternative way.

But there are some constrains that should be on developers and their code. As l0b0 put it the code should be modular, testable and readable. For example, Nvidia can provide an opensource driver to their cards, but it can be unreadable, it can be intentially obfuscared. So, this "opensource" drivers will be unuseful and even harmful for community. Because there's acctually no difference between proprietary code and obfuscared code.