Long Peaks

Anyone looked a the front page?
So is Long Peaks going to be only for SM 3.0 hw and beyond?

http://www.theinquirer.net/default.aspx?article=39846

So is Long Peaks going to be only for SM 3.0 hw and beyond?
I see nothing in the article about any specific hardware requirements for LP, (which they feel will be called “OpenGL 2.X”) but everything we’ve been led to believe promotes the following requirements:

  • Shader-aware hardware. R300/NV30 or better.
  • FBO-capable hardware. Same as above.
  • VBO-capable hardware. Pretty much anything.

Mt Evans (apparently, the article thinks it will be called "OpenGL 3.0) will have the stricter hardware requirements.

The article isn’t a very good one. It doesn’t even mention the fact that those DX10 features are already available in OpenGL now through extensions.

Indeed, the only interesting thing in that article are the new names: OpenGL 2.x and OpenGL 3.0. Which I find to be… wierd.

I mean, I understand the principle of the argument. Longs Peak, while a major change to the user of OpenGL, ultimately provides no greater power for anyone using an LP-capable GL application. Whereas Mt. Evans, being about features more than API, will provide a substantial increase in visual quality for anyone using a Mt. Evans-capable GL app.

A GL 3.0 application compared to a GL 2.x sounds much more advanced than a GL 3.1 application compared to a GL 3.0. As such, the version name for LP is diminutive compared to Mt. Evans.

Even so, it’s entirely inappropriate for users of OpenGL. For us, 2.x is a major change, while 3.0 is just some new functions.

I think they should rename the api to OpenGI, the ‘I’ standing for Interface rather than library.

I think Mr. Leech’s words were that LP was just a recasting of 2.0. There’s been no mention of specific hardware requirements that I’m aware of.

Now for Mt. Evans, I think it probably reasonable to assume a SM4 hardware requirement, which for most of the extensions we’re talking about is guaranteed by D3D10. But I think SM3 would be difficult to guarantee at this point.

P.S. V-man, not to put too fine a point on it, but I think there’s just one peak, and it’s Long(’)s. Long Peaks sounds like a daytime drama :wink:

Well, I hope that Long Peaks (I am interested in the Lean and mean profile) will be available on all SM 2.0 hw.

So what exactly is added in Evans?
There is some stuff about long peaks but I haven’t seen anything about Evans.

http://www.opengl.org/pipeline/vol003/
“Using the Longs Peak Object Model”

V-Man, it’s all there in Volume 3: http://www.opengl.org/pipeline/article/vol003_5/

OpenGL 2.x and OpenGL 3.0. Which I find to be… wierd.
It makes sense from a marketing point of view. Then you can say “OpenGL 3.0 hardware” instead of “DX10 hardware”. And this sounds a lot better than “OpenGL 3.1 hardware” :wink:

they really did shoot themselves in the foot by using up 2.0 when they did. The 2.0 label should have been reserved for a complete API refactor such as this, not just for a flurry of handy extensions.
Still, hindsight is 20/20 as they say.
The API definitely needs a completely new name - it just isn’t OpenGL any more. I don’t see what difference it will make from a marketing standpoint - it’s only really developers that know or care about OpenGL, and they will follow the change in name quite easily.

I’m not a marketing specialist, but I think it’s easier to sell a “better version” of OpenGL than something completely new. It doesn’t have anything to do with what it really is. Marketing rarely does have anything to do with what it really is :wink:

It may not be the old OpenGL, but it will be ‘the’ OpenGL. They’d be silly to throw away the OpenGL name.

Personally I think Longs Peak should be OpenGL 3.0 and Mount Evans should be 3.1.

Regards
elFarto

personally, given the ‘source’ of the information seems to be The Inquirer I’m not putting too much faith in it…

OpenGI?
Gee… :wink:

When I first heard of 2.0 I expected some serious API changes - perhaps not as much as Longs Peak will bring but still many. At least a proper render to texture functionality. Note that FBO’s arrived as extension some time after OpenGL 2.0.

In current situation I think 2.x and 3.0 are just fine. I simply agree with Overmind’s point of view. As developers we really shoudn’t care.

for me, opengl now has a bad reputation - no longer something to be proud of. It’s renowned for its clumsy, overly complicated, outdated and non-standardised extension mess. When 2.0 came out it was clear that the version number was no indicator of progress, and people will think the same of 3.0.
We all know that longs peak will be a completely new API, but everybody else will assume it’s just another minor update.

To hell with the naysayers!

“OpenGL is back, and this API is all piss and vinegar.”
– The Frugal API

“OpenGL 3.0 opens a whole new six-pack of API kick ass.”
– Graphics Practitioner’s Weekly

“OpenGL 3.0 marks the return of the greatest graphics API the world has ever known.”
– Patron, Joe’s Bar and Grill

“Two thumbs up.”
– Eguart and Raster

I’m with k_szczech and Overmind. Still I think one shouldn’t drop the codenames. I mean “OpenGL 2.x” doesn’t sound as nice as “OpenGL 2.x Longs Peak” :smiley:

I don’t really care, you know. I’m just procrastinating over my own stuff again. We can milk this discussion a bit more, can’t we?

Originally posted by knackered:
for me, opengl now has a bad reputation - no longer something to be proud of. It’s renowned for its clumsy, overly complicated, outdated and non-standardised extension mess. When 2.0 came out it was clear that the version number was no indicator of progress, and people will think the same of 3.0.
We all know that longs peak will be a completely new API, but everybody else will assume it’s just another minor update.

“Everybody else” doesn’t know what OpenGL is, at all. It’s only us developers, who actually care, what 2.0, 2.x, 3.0, … really means. For everybody else it’s just, “hey my card supports OpenGL 3.0, great!” and then “what is OpenGL anyway?” (actually i didn’t make this up, i really overheard someone saying that, if you substitute OpenGL with TnL).

As a developer one thing you usually learn early, is to ignore PR-stuff. And renaming OpenGL into OpenGI or whatever would be just that, PR. It does not change a thing for the people who actually make use of it.

Although i agree, that Long’s Peak should be 3.0.

Jan.

“although I reserve the right to contradict myself at the end of my speech”

I think Longs Peak should be OpenGL 2.5 and Mount Evans should be 3.0 : why ?

  • same features as 2.1, heavy changes in the API : increment the minor version number a lot.
  • new features/new hardware : increment the major version number.

I think if GL can make describing a video card as easy as DX does, it would be finally worth it to mention the version number.

“Hey, I have a GL 2.0 card”
“Hey I have a GL 3.0 card”

Right now, from Radeon 9500 to the X2900 and from Geforce FX 5200 to Geforce 8800, they are all GL 2.1 cards, even though some of those don’t have GL 2.1 drivers due to laziness from the IHV.

GL 2.1 covers 3 generation of shader model.

Meanwhile, in the D3D world. There is SM 2.0 (is this Dx9A?)
SM 2.x (I think DX9b)
SM 3.0 (DX9c)
SM 4.0 (DX10)