Re: [Glgui-dev] Issues

[Scott F. <sf...@gl...>]

> I suggest a fully implemented Bridge.  We can then allow any developers
who
> dislike our implementations of the D3D and OpenGL classes extend the
> abstract base class which defines the interface easily.  Or create a
simple
> adapter and integrate their own versions of D3D OpenGL bridge with our
own.
> That would also allow a very easy extension to anything else, even (dare I
> say it) a software renderer.  Then no ports will be necessary, we can just
> derive from GraphicsAPI (or whatever we want to call it.)  This is what I
do
> in my engine, and it is one heck of a timesaver.  Admittedly it will add
an
> extra layer, but I think the increased flexibility would be worth it.

This was what I was going to suggest.  As of now, there's a pretty limited
portion of the GL that needs to be abstracted.  I haven't used d3d in a long
time, and don't know how other api's implement similar functionality, so
I'll leave it to someone else to lay out the abstraction class.  But I'll
write the opengl version once the base interface is done.

> Also let me say here and now, I love the stl.  I know on a few systems it
is
> still unsupported, but it is the biggest time saver save templates
> themselves.  I say we should use it to its fullest.

I like it too, but if there are too many support problems, it may not be
worth using.

> PS. Sorry I haven't been active, school has hit me harder than I thought
it
> would.  This is finals week, so hopeful If I survive I should be able to
get
> back into the swing of things next week.

Same goes for me.  Except that I'm in my last week of classes and will be
starting finals next week.  I'm also lining up a job for the summer which
will include a good bit of time moving, etc.  I may be unable to do much
work for a few weeks still.  But I'll try to stay a part of the discussion
and can even spend a few hours here and there to do some simple design /
programming.

One of the things I'd like to work on next is a more complete shader class,
with functionality similar to that in quake3.  But the issues Keith has
made, as well as some of the others that have been mentioned (other widgets,
better text, etc) should take precedence.

Scott Franke [druid-]
   sf...@gl...
druid-'s GL Journal
   http://www.gljournal.org