Re: OpenGL+GLUT the future

[Paul Matthews <paulmatthews@yahoo.com>]

---Ian Crawford <icrawfor@uoguelph.ca> wrote:
>
>> The obvious conclusion of these two trains of
>> thoughts is that the best way games development, 
>> and playing, on Linux can progress is by improving
>> Mesa and by expanding GLUT.
> 
> Interesting opinion.  Please don't be offended that
> I don't agree. I'll say right now that I don't like 
> GLUT.  I feel that it abstracts too much from the 
> user.  It's great for demos.  I mean, little sample
> programs that will run on any platform are a great
> means to learn any type of programming, but I don't 
> think it's cut out for tomorrows games-to-be.  Is 
> there even a glide version of GLUT?
> 
It's abstraction is not a disadvantage, its an advantage. And it's not
inefficient, look at the source code. You cant get much learner and
more efficent than that.

Glut is built on top of Mesa/OpenGL. If Mesa/OpenGL has glide support
(yes it does) then GLUT has glide support.

Sounds like you dont know much about GLUT.

>> GLUT 3.7 already handles joysticks, spaceballs and
>> the keyboards in a platform independant manner, why
>> not add platform independant sound support to GLUT
>> (PenguinSound)?
> 
> GLUT needs more than that.  There's also networking,
> cut scenes, native support for common 3D model
> formats...  On another note, is there an object 
> oriented version of GLUT?  I just don't think it's a
> good basis for game development.
> 
GLUT is GPLed. And it's author is expanding its library to support
game programming. All of that mentioned above could intergraded into
GLUT, or form an add on library. If want networking etc, why not make
it platform independant? 

>> PenguinGUI can be a layer over GLUT the same way MUI
>> is in GLUT already.
>
> I think that's building up too much.  We're starting
> to look at some big performance digs.
>
Not really, a Messa/GLUT/GUI can be faster than X11, simply because it
can take advantage of the hardware in a way X11 cannot. If also allows
for cool alpha effects etc all.

>> And Brian Paul could certainly do with a hand making
>> Mesa faster...
> 
> I agree, but a software renderer can only do so much.
> What mesa really needs is good hardware support for 
> a wide variety of hardware, and this is a work in 
> progress (yay!).
> 

There is a large scope for improvement in the software renderer, and
geometry engine. 

But Mesa gives us the doorway to the current and next generation of 3d
and geometry cards, a far better way to go than assuming that the
hardware can only "draw", like direct3d (and penguinGraphics(???)) does.

Do you think the current high end cards are optimised for OpenGL,
direct3d or penguinplayGraphics?

--
Paul Matthews
==
--
Paul Matthews
nudge@samba.anu.edu.au


_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com