What does everyone think of 2.0?

I was just wondering how everyone feels about the 2.0 proposal. I looked over the spec and the only complaint I have is that the spec makes me want to re-implement half of my game engine. I see allot of flexibility in the new architecture, however my experience is limited to game development and I was wondering how the rest of the community is reacting.

Thanks…

John.

I like it very much, on the whole. I would like to see a basic framework for extending vertex programs to one-in-many-out scenarios, and I still can’t see how to get displacement mapping within ten miles of that pipeline, but generally I think they’re making all the right choices.

One thing that does bother me is that GL2 will be SOOOO far removed from 1.x. Realistically, I can’t see any cards fully supporting GL2 in hardware for a couple more generations, and by that time there’ll be a huge installed base who won’t be too keen on throwing out that GF4 they just bought. Trouble is, the APIs are so different that supporting both could end up requiring two engines and two sets of art. For the gaming market at least I guess this’ll end up making the highend/lowend divide even wider, to the point where effectively different platforms.

Hey I know this is not the place but there is an interesting thread in the Advanced Coding board. Here is the link every one is talking about…
http://www.theregister.co.uk/content/54/23708.html

I do like the ideas currently in the 2.0 spec. But only one problem is the name “shader,” as it has already been pointed out on a number of occasions, it should be “program.”

Just a something I thought needed pointing out :wink:

Also, I’d like to know:

  1. The estimated timescale of a 2.0 release?
  2. When the OSS OpenGL SI is going to be updated?
  3. Is Brian Paul getting ready for a new version of Mesa? BTW, wasn’t Mesa & OpenGL SI going to be combined in the future as a replacement for the SGI SI?

Thanks,
Luke.

I also like “vertex program” more, because it actually more precise, but I guess they have chosen “shader” for marketing reasons.

-Lev

Just out of curiosity what are you guys using OpenGL for? I was just wondering how much of the computer industry uses OpenGL? For example, I don’t use OpenGL at my job, but I’ve been working on a game engine at home in my spare time for a few years now. I’ve spent a lot of time on OpenGL.org, but I’ve only just started reading the posts and I’ve noticed that a lot of people post very regularly. Are all these people using OpenGL at work or at home, for game development or graphic design, or just for fun?

John.

Originally posted by john_at_kbs_is:
[b]Just out of curiosity what are you guys using OpenGL for? I was just wondering how much of the computer industry uses OpenGL? For example, I don’t use OpenGL at my job, but I’ve been working on a game engine at home in my spare time for a few years now. I’ve spent a lot of time on OpenGL.org, but I’ve only just started reading the posts and I’ve noticed that a lot of people post very regularly. Are all these people using OpenGL at work or at home, for game development or graphic design, or just for fun?

John.[/b]

I would imagine most poeple reply in the advanced developer forum are developing software relating to OpenGL.
A lot of people posting questions in beginner forum are doing it for fun, or for a school project.
Keep in mind almost all professional 3D applications are written in OpenGL, that means a lot of people are making their livings by writting applications with OpenGL.

Originally posted by john_at_kbs_is:
[b]Just out of curiosity what are you guys using OpenGL for? I was just wondering how much of the computer industry uses OpenGL? For example, I don’t use OpenGL at my job, but I’ve been working on a game engine at home in my spare time for a few years now. I’ve spent a lot of time on OpenGL.org, but I’ve only just started reading the posts and I’ve noticed that a lot of people post very regularly. Are all these people using OpenGL at work or at home, for game development or graphic design, or just for fun?

John.[/b]

I developed a GL driver for a 3D engine for work - they weren’t too bothered, but I wanted to do it for cross platform reasons. Plus it made the engine more portable - cuz it wasn’t at the time

Luke.

P.S: My 3D knowledge is next to nothing and my GL knowledge is below what I’d like it to be. How amusing is that?

I think that OGL2 specs proposal is really interesting but very complicate for every poeple that don’t have the knowledge of the 3dlabs team members !
For me, the best API is an API that permit to do simplest thinks in the most immediate manner, and that offer more than one level of use. So, simplify, unify, make 2 or 3 levels of use with same API (no 2 or 3 API), make a more generic methods for basics features, integrate more processing between the OGL2 API and the drivers and all will be the best for me !

Gaby

Originally posted by gaby:
[b]I think that OGL2 specs proposal is really interesting but very complicate for every poeple that don’t have the knowledge of the 3dlabs team members !
For me, the best API is an API that permit to do simplest thinks in the most immediate manner, and that offer more than one level of use. So, simplify, unify, make 2 or 3 levels of use with same API (no 2 or 3 API), make a more generic methods for basics features, integrate more processing between the OGL2 API and the drivers and all will be the best for me !

Gaby[/b]

Well, this is what they are suggesting. An intermediate step; OpenGL 2.0 which includes the new programmable API’s.

The next release would be pure OpenGL 2.0 which is just the programmable API’s really.

From what I’ve seen, it looks like a good spec for the next gen GL.

I’m looking forward to playing with it.

Luke.

Just a quick note on the term ‘shader’ and ‘program’. Our latest thinking is as follows:

Similar to texture objects, that are containers for image objects, we’ve defined program objects. These program objects are containers that hold shader objects. A shader object holds some amount of source code, not necessarily a whole program. You have to make an API call to compile a shader.

Multiple shader objects in one program object get linked together, and produce an executable that gets loaded into one (or more) of the OpenGL 2.0 defined Processors. Currently there are the vertex processor, fragment processor, unpack processor, and pack processor.

Thus you can have n number of shaders, each with some of your source code in it to form one program that gets run. It really is very similar to modules that get compiled, and later linked together to form an executable, when using your favorite development environment.

Barthold
3Dlabs

— OFFENSIVE POSTING WARNING !!! —
Hey Gaby,
I read your postings and it seems, that you never developed a real application like a game or visualizion system. I’ve the feeling, that you just saw some lines of code or read the first 2 Chapters of the Spec. If you’d rellay had programmed a real app using OpenGL then you’d know, that it’s the most simple API you can imagine and that OGL2 is just the logical step to provide the same flexibility for the new features as known from OGL1. Of course I may be wrong, but then please prove it. A short but good code snipet is enough. Post it here or send it to wdraxinger@darkstargames.de please…

[This message has been edited by heXAriTH (edited 02-06-2002).]

Well, you know, I said what I feel in my position… What I’m saying is not the truth, it’s only my opinion !

Yes, I was game developper, and I expected coming back in this industry in next year, with a big project… I start to learn OpenGL for two years ago only, most in my spare time. In the past, in 1995, I’ve done a racing game for Kalisto (Al Unzer Jr Arcade Racing) : I’ve done 80 % of the game (coding, graphics). It was a software rasterizer. After this, i’ve coded a classic software 3d engine.

In 1998, I’ve made a Racing game for Michelin (tire manufacturer), in six month, for wich I’ve build a modeler to create all the meshs and edit game datas in the most efficient manner. You can see it here at www.z-oxyde.com. This game was based on Glide. After this I’ve started a big game developpement environment project based on OpenGL : on this basis, i’ve made models for Michelin F1 web-site.
…etc…
So, if you go on our site web, remember that I’ve made 80% of all what is on this site (CG, editing tools, proessionnal software). So, you will understand that I’m not using OGL all time. I hoppe it will provide me the simplest manner to reach the most advanced features of hardware.

I know that I should use higher level interface provided by 3d engines like renderware, but I think it’s a good thing to take care of people wich cannot spend more time on too complex low level API. It’s a manner to avoid the design of APIs used only by people able to make it. That’s what I want to say !

PS : pardon for my bad english !

Gaby

<OFFENSIVE CONTENT>

heXAriTH:

  1. Open mouth.
  2. Insert foot.

</OFFENSIVE CONTENT>

j

heXAriTH :

I’ve seen that due to the point at end of the link it doesn’t work.

So try this : www.z-oxyde.com

You can see this, wich is a real-time demo in OGL of a simple “advanced” realtime squeleton with muscle principle :
http://www.z-oxyde.com/research/

Thank’s ! Keep cool, be efficient in your point of views !

Gaby

gaby:

I’ve to apologize at you. My opinion was, that there were a few lines in your postings that seemes obsolete or wrong. Especially your statement, that only the people at 3Dlabs relly understand their own shaders…
I’m now programming OpenGL since 1999 ( you see, not much longer than you ;-), however I’m still to gymnasium (german high school) so I got a lot of spare time, coding every time, sorry I lost the thread ;-). I think, that the new ideas on OGL2 are really cool, and since I’m a LINUX guy a major version number increase means to me also the abolishment of old unneccesary features. OK, there’s the big question on DLs, I’m completely in their favor, but only for state management. Its simpler to do a glCallList(STATE_FOR_SKY_RENDERING); or what else, then to do a glDisable(GL_LIGHTING); glDisable(GL_DEPTH_TEST); …
But the real part of OGL2 catched me, and I think not a lot should be changed, since the 3Dlabs guy did in my eyes a tough job.

<TOO OFFENSIVE CONTENT>

j:

  1. umount /dev/hda1
  2. dd if=/dev/zero of=/dev/hda1

</TOO OFFENSIVE CONTENT>

Originally posted by john_at_kbs_is:
I was just wondering how everyone feels about the 2.0 proposal.

I find this a bit cruel… I mean, the specs are great, but it’s so hard to wait until it finally gets released (

I need unified vertex shaders, I need auxilliary buffers writeable from vertex shaders… I want it all… I want it all… I want it now !

(Anyway, I got these old Queen CDs to help me waiting )

Julien.

Originally posted by heXAriTH:
OK, there’s the big question on DLs, I’m completely in their favor, but only for state management.

AFAIK, DL are used in distributed rendering applications: the guys using them may count for 80% (evil assumption) of the total number of OGL developers.

Julien.

I love it. I would love it even more if I could actually use it. Let’s go OpenGL 2.0! Lets go lets go!