An OpenGL 4.1 Gallium3D State Tracker Was Just Proposed

[V!NCENT]

I would also like to say that I understand that it affects the US citizens and that we should keep the entire world in mind, but likewise nobody cares for chinese laws. Not to forget: China has more people than the US.

Let us remind ourselves that we should not always accept any law just because it's the law. If you could get punished for not handing jews to the Nazi's then does that mean that you should? Why is the free software movement limited by the law to compete against a giant software company that is or was, by law, anticompetative? Something is realy wrong here!

[89c51]

Let us remind ourselves that we should not always accept any law just because it's the law. If you could get punished for not handing jews to the Nazi's then does that mean that you should? Why is the free software movement limited by the law to compete against a giant software company that is or was, by law, anticompetative? Something is realy wrong here!

The companies making money on open source are affected by the patents not the community. We have code for all (i think) the patented stuff (from h264 to whatever) and anyone can activate these features (use a different repo or whatever the distribution has).

What harms open source is the lack of manpower and the fragmented nature of it. (ie they just forked ffmpeg).

[steckdenis]

Hello,

Thanks to everyone for the encouragement. I will now give you some more information about my project.

At first, it's a project. I really hope to have something interresting done during the summer, but I cannot say for sure that I will succeed to make a complete Open GL 4.1 + OpenCL + OpenGL ES 2 state tracker in three months . As said on this forum and on the Mesa mailing list, it's a really big adventure and my GSoC project is to begin it and then to go as far as possible.

Many people are also concerned about the software patents. Most of OpenGL is free of patents (the features covered by OpenGL 2.1 and big parts of OpenGL 3+). The most annoying patent for the moment is the one covering floating-point textures. Fortunately, it doesn't cover software implementations, so the state tracker and llvmpipe can use it. I don't think that the fact I work in Belgium changes the situation regarding the patents. Mesa is used in the entire world, including the USA. One way to go could be to use the --enable-patented switch when building Mesa. One inconvenient of this is that we have to rebuild Mesa to enable patented features, but a distribution based out of the USA could provide two Mesa packages, one for the US people, and another for people living outside the USA.

I also want to say that I don't do this work for money. I follow Mesa for two years and I want to do that for nearly one year. The Google Summer of Code is only a big plus that decided me to try the adventure, but I will also work on this state tracker even if I'm not selected by Google, and I will begin when I have time and when other Mesa developers are ok if Google allows that (I mean that I will not wait until the summer and then stop just after the Summer of Code program).

Thanks to everyone,
Denis Steckelmacher.

[89c51]

you've probably read corbin's and marek's answer in the mailing list and i think they are right.


Do something small,fast and usefull (ie parts of the GL todo list). No need to aim really high.

[elanthis]

I would also like to say that I understand that it affects the US citizens and that we should keep the entire world in mind, but likewise nobody cares for chinese laws. Not to forget: China has more people than the US.

The US has more people with cash and more people who are thoughtless consumers, though. We buy and use more shit than anymore else. An entire nation of millions of people who consume 100x as many tonnes of product as they produce.

Let us remind ourselves that we should not always accept any law just because it's the law.

Let's also remind ourselves that getting sued out of existence isn't in any way fighting a law. Protesting against a law in mass is effective. Being downtrodden outside of the wider public eye accomplishes nothing. And the common man doesn't know his ass from a hole in the ground when it comes to patents or other "IP" laws. He's not going to get behind protesting on your behalf if you get sued for patent infringement. You'll get a small handful of uber-nerd geeks sending a handful of emails and that'd be the extent of the support you're going to get.

[drag]

Sure it's possible. Do all your work in the European Union. They don't have the retarded concept of software patents.

What I find irritating is people that say shit like this without much clue about what they are talking about.

Saying shit like 'Oh Software patents are a USA only problem' is ignorance of the Nth degree. It's just a stupid wrong thing to say.

What makes it worse is a lot of these people are from Europe, who you would think would actually now more about their own laws then USA ones... unfortunately most of them seem that they do not.

[Veerappan]

While we eventually would like the OpenGL 3/4 and OpenCL state trackers to all work wonderfully, I'd suggest picking either the OpenGL 3 or OpenCL state tracker to begin with and see how far you get.

I would personally be overjoyed if the clover project was finished by someone, and in the future if I am cursed with an overabundance of free time, I might take a stab at it myself. I wouldn't be disappointed if someone else beat me to it, however.

[drag]

you've probably read corbin's and marek's answer in the mailing list and i think they are right.


Do something small,fast and usefull (ie parts of the GL todo list). No need to aim really high.

Maybe he is so brilliant that he IS aiming low. Geniuses do exist, although they are exceedingly rare.

My only suggestion is that if he wants to be taken seriously, as he must first convince the developers to take him on, he must do something to prove himself.

[steckdenis]

I agree. I will in some days try to begin a "proof-of-concept" (but I hope it will be useful for the final state tracker, if any). I want to experiment with the context creation and the GLSL-compiler. It will already take me a fair amount of time to do that, but I will be able to see if it is possible.

[Veerappan]

Yeah, I'd recommend starting sooner rather than later. Even proof of concepts and prototypes can be immensely useful when determining what paths you want to go down.

I would like to paraphrase an unknown source from my thesis-writing experience:

The best thesis projects are those that are 90% completed before the proposal is approved