New X.Org Release Process Has Been Reached

[phoronix]

Phoronix: New X.Org Release Process Has Been Reached

Last week we talked about a new X.Org release process proposal for improving the consistency and quality of X Server releases through taking a number of relatively simple steps. Well, this week from XDS2009, a revised proposal has been agreed upon now making it policy for X Server 1.8 / X.Org 7.6 and later. With this new process, there will be consistent six-month releases that should be very predictable...

http://www.phoronix.com/vr.php?view=NzU3Ng

[perpetualrabbit]

I would like to start a discussion about linux graphics progress:

Is the progress really that slow? (It seems to me compared to OS X and windows)
Would it be possible for someone with a couple of millions to spend to simply step in and hire some programmers to work on key projects, like a free Nvidia driver (or help nouveau out), the Xorg server etc.?
Or is the problem that those kind of programmers cannot be found (i.e. too hard, or too much knowledge needed to get into Xorg programming), in other words, everyone who is able to work on say Xorg already works on it?

So in short, is it money, is it people, is it both?
Or is it lack of leadership and focus?

If it is NOT money, then why don't the ones who have it (Mark Shuttleworth, Redhat) start a high priority project to bring linux graphics up to OS X standards, or Windows 7 standards (without the bloat), and quickly (say 1 or 2 years)?

If the people who can do it are there, what is stopping them?

[whizse]

Sitting around and waiting for a millionaire to show up and fund development seems like a really horrible idea.

Red Hat, and other companies, does hire people to hack on X and related projects. I guess you could say that the work they, and companies like Intel, Tungsten etc. are doing really is the big project to bring free graphics up to contemporary standards.

[perpetualrabbit]

* Intel, IBM, Canonical and others have money. A couple of million is nothing for them. Also, they have a vested interest in Linux, and would reap profits if the linux graphics stack is on par with OS X or windows 7.
* Yet they employ only a handful of developers working on Xorg, drivers, KMS and other linux graphics subsystems.
* Just a wild estimate: the real hard work on those subsystems seems to be done by about 20 or 30 people right now. The hardcore coders like Keith Packard.
* At the current rate, it looks to me it will be 5 years before linux graphics is on par with OS X or windows 7.
* A really good graphics subsystem is needed _now_ to compete with OS X or windows. I name a few field to compete in: 1)graphics design and photography, 2)Gaming, 3)Media center. Missing in X window system for
1) Color profiling, higher than 8 bits per color channel
2) Free, hardware accelerated OpenGL 3.2 for at least nvidia, ati and intel. A compelling development environment for games.
3) Free implementation of vdpau for all cards that have video decode chips. Video overlays working with 3D desktop.
* I would advance $200 for a (open source) great graphics subsystem. If ten thousand other linux enthousiasts can be found to shell out $200 in advance, your have two million. For two million you can hire 20 good developers for one year. Assuming that they can be found, you have a good development plan and clear goals.


I really think devepment is too slow right now, and in five years time it will be too late for linux to compete.

I just spend another evening without success to get vdpau working, and countless times had a text-only screen after a reboot because the kernel was updated and I use the latest driver downloaded from nvidia.
$200 for a no-worries always working graphics system seems worth it.

[unimatrix]

I completely agree with perpetualrabbit and I would also gladly give $200 to get this whole video mess move faster. 9998 to go.
We should make some kind of a web directory where people could sign up if (and how much) they're willing to pay. Just to get a rough estimate of how much we could accumulate.
Of course the question "where to donate that money" immediately arises.

[Apopas]

I have thik about this as well many times. Something similar happened in 2004 and mozilla-foundation collected 250,000$ and advertised Firefox in NYtimes and before when 100,000$ collected so the community to buy Blender and make it opensource.
Now, Firefox is a really competing browser with a large market share. Over IE7-8, over Safari and behind IE6. With a different point of view it's similar as if Linux is over Vista, Win7 and Mac but behind WinXP, a very fortunate conclusion.
But to collect money so to be able later to pay the developers' salaries is a totally different thing. A company should get the responsibility and yet how could everyone trust a company to give them his money. I don't say isn't applicable but for sure is more difficult.

[d2kx]

I think the development is faster than it ever was before right now, but that it is still too slow. I mean, the Linux desktop has been promoted for years and it lacks proper support of simple features if you describe the current situation in the dramatic way. I also am wondering why companies are not hiring more developers.

[Bryce Harrington]

* Intel, IBM, Canonical and others have money. A couple of million is nothing for them. ...
So in short, is it money, is it people, is it both?

We're short on billionaires actually. Please go become a billionaire, or perhaps convince one of your friends to become one. It can't be that hard, I'd be one it if I only had the time. Don't worry about deciding how to spend the money yourself, the community will give you directions for that.

[Apopas]

We're short on billionaires actually. Please go become a billionaire, or perhaps convince one of your friends to become one. It can't be that hard, I'd be one it if I only had the time. Don't worry about deciding how to spend the money yourself, the community will give you directions for that.

No need to be rude. The guy just said that if a company lives from Linux and they have money, it would be an investment if they hired more developers for X.

[msebast]

It's been tried before. Giving developers money is harder then you would think.
http://www.pledgebank.com/nouveaudriver?showall=1

Anyone looking to collect money from end users and give it to developers should keep that attempt in mind. I signed that petition but never got to send any money. The Nouveau developers seemed rather ambivalent about the whole thing.

Throwing millions of dollars around or hiring hundreds of developers does not necessarily result in faster delivery or high quality code. (See mythical man month, FAA air traffic control, etc.)

The current rate of progress is remarkable. The Phoronix news sidebar shows a continuous stream of improvements. The right people are working on this and they've been hired by the right companies. (Redhat, Intel, AMD, Tungsten, etc.) Those companies handle the task of collecting money from end users and giving it to the developers.

We'd all like to see faster progress but doing things right takes time. They've basically ripped the entire Linux graphics system apart and are re-architecting everything. But they are doing this carefully so the software is still usable by the distros and there have been few notable regressions. (like Intel in the last Ubuntu version.)

I think your concerns about meeting a five year market window are misplaced. The beauty of open source is that it does not need to meet a market window. Developers take their time and make the software right. I'd rather have working, stable software next year then broken, buggy crap today.