Re: [Haskell] ANNOUNCE: OpenGLRaw 1.0.0.0

On Thu, Jun 11, 2009 at 06:56:01PM +0200, Sven Panne wrote:
Version 1.0.0.0 covers the OpenGL 3.1 core, all ARB extensions and all EXT extensions.
What about OpenGL 2.1? As I understand, Linux won't have OpenGL 3.0 or 3.1 for at least some months, $ glxinfo | grep 'OpenGL version' OpenGL version string: 2.1 Mesa 7.5-rc3 Thanks! -- Felipe.

On Thu, Jun 11, 2009 at 19:37, Felipe Lessa
On Thu, Jun 11, 2009 at 06:56:01PM +0200, Sven Panne wrote:
Version 1.0.0.0 covers the OpenGL 3.1 core, all ARB extensions and all EXT extensions.
What about OpenGL 2.1? As I understand, Linux won't have OpenGL 3.0 or 3.1 for at least some months,
$ glxinfo | grep 'OpenGL version' OpenGL version string: 2.1 Mesa 7.5-rc3
On my machine I get: $ glxinfo | grep -i version server glx version string: 1.4 client glx version string: 1.4 GLX version: 1.3 OpenGL version string: 3.0.0 NVIDIA 180.51 OpenGL shading language version string: 1.30 NVIDIA via Cg compiler So it doesn't appear to be true. Best regards Krzysztof Skrzętnicki

On Thu, Jun 11, 2009 at 07:55:23PM +0200, Krzysztof Skrzętnicki wrote:
On my machine I get:
$ glxinfo | grep -i version server glx version string: 1.4 client glx version string: 1.4 GLX version: 1.3 OpenGL version string: 3.0.0 NVIDIA 180.51 OpenGL shading language version string: 1.30 NVIDIA via Cg compiler
So it doesn't appear to be true.
So, open source Linux drivers won't have OpenGL 3.0 or 3.1 for at least some months :). I don't know the internals of the graphics subsystem, but AFAIK the open source drivers use Mesa to implement the OpenGL frontend, and the current version 7.5rc3 doesn't support it yet. FWIW, I'm using the Intel driver, version 2.7.1. -- Felipe.

nVidia and ATI drivers both support GL 3.0 on Linux, although you're
right that open source drivers don't. I for one welcome this package
with open arms, since I'm mostly trying to implement a layer over
OpenGL anyway with Hieroglyph. This'll help with the next revision of
that. As for the package naming scheme, I don't mind it at all --
there are several different Graphics.Rendering packages with different
backends and several Graphics.UI packages, so I think it's been fairly
established.
Perhaps there should be a Hackage Ontology project, though, or we
should ask a librarian to look at Hackage and think about the problem
and suggest a proposed standard for people to follow, then suggest to
new library releasers under what set of QNames they ought to release
their libraries under. This is more or less standard practice for
large working groups on languages like XML and Java, and although it
could be seen as a debatable practice to many, I prefer something
descriptive of function in my module names, since package names so
rarely are.
case in point: Hieroglyph. What's it do? import Hieroglyph. Is
there any clue by my function names which ones belong to a library
called Hieroglyph? No. However, import
Graphics.Rendering.Hieroglyph, and I see a function somewhere in the
code called "arc" or "plane" or "circle", and I know it probably goes
with the rendering package.
On Thu, Jun 11, 2009 at 2:14 PM, Felipe Lessa
On Thu, Jun 11, 2009 at 07:55:23PM +0200, Krzysztof Skrzętnicki wrote:
On my machine I get:
$ glxinfo | grep -i version server glx version string: 1.4 client glx version string: 1.4 GLX version: 1.3 OpenGL version string: 3.0.0 NVIDIA 180.51 OpenGL shading language version string: 1.30 NVIDIA via Cg compiler
So it doesn't appear to be true.
So, open source Linux drivers won't have OpenGL 3.0 or 3.1 for at least some months :). I don't know the internals of the graphics subsystem, but AFAIK the open source drivers use Mesa to implement the OpenGL frontend, and the current version 7.5rc3 doesn't support it yet.
FWIW, I'm using the Intel driver, version 2.7.1.
-- Felipe. _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe

On Thu, 11 Jun 2009, Jeff Heard wrote:
case in point: Hieroglyph. What's it do? import Hieroglyph. Is there any clue by my function names which ones belong to a library called Hieroglyph? No. However, import Graphics.Rendering.Hieroglyph, and I see a function somewhere in the code called "arc" or "plane" or "circle", and I know it probably goes with the rendering package.

Oh, and I don't disagree with that at all. I just just have an
aesthetic preference for multiply qualified library names. Chalk it
up to the fact that my partner's a librarian, so I'm used to putting
things in categories, subcategories, and sub-sub-categories :-)
-- Jeff
On Thu, Jun 11, 2009 at 2:44 PM, Henning
Thielemann
On Thu, 11 Jun 2009, Jeff Heard wrote:
case in point: Hieroglyph. What's it do? import Hieroglyph. Is there any clue by my function names which ones belong to a library called Hieroglyph? No. However, import Graphics.Rendering.Hieroglyph, and I see a function somewhere in the code called "arc" or "plane" or "circle", and I know it probably goes with the rendering package.

And the one other thing is that it increases (to me) the at-a-glance
comprehensibility of the module. If I'm reading over soemone else's
code and I want to get a feel for where s/he put things, the fully
qualified module structure and the fully qualified names in the import
statements make it easy to get a feel for the organizational structure
of the program.
That way, on first reading I don't have to go through the whole module
to figure out what was contained in Utils when I see the phrase
import Utils
or even
import qualified Utils
at the top of the module.
-- Jeff
On Thu, Jun 11, 2009 at 2:55 PM, Jeff Heard
Oh, and I don't disagree with that at all. I just just have an aesthetic preference for multiply qualified library names. Chalk it up to the fact that my partner's a librarian, so I'm used to putting things in categories, subcategories, and sub-sub-categories :-)
-- Jeff
On Thu, Jun 11, 2009 at 2:44 PM, Henning Thielemann
wrote: On Thu, 11 Jun 2009, Jeff Heard wrote:
case in point: Hieroglyph. What's it do? import Hieroglyph. Is there any clue by my function names which ones belong to a library called Hieroglyph? No. However, import Graphics.Rendering.Hieroglyph, and I see a function somewhere in the code called "arc" or "plane" or "circle", and I know it probably goes with the rendering package.

On Thu, Jun 11, 2009 at 11:55 AM, Jeff Heard
Oh, and I don't disagree with that at all. I just just have an aesthetic preference for multiply qualified library names. Chalk it up to the fact that my partner's a librarian, so I'm used to putting things in categories, subcategories, and sub-sub-categories :-)
Check out http://www.shirky.com/writings/ontology_overrated.html .
participants (5)
-
Conal Elliott
-
Felipe Lessa
-
Henning Thielemann
-
Jeff Heard
-
Krzysztof Skrzętnicki