On Wed, 9 Jan 2002, Rob Healey wrote:

> 	I believe Jason mentioned a good course of action would be to get
> 	the gtk librarys working down on the bare metal rather than having
> 	X in the way.
> 
> 	The dreamcast doesn't have enough memory to use a memory pig like X
> 	as the graphics system, 16M and X just don't get along well...

  Actually, I'm told on multiple fronts that the "X is bloated" thing is
largely a myth - listing memory usage counts the memory apertures to the
graphics card's frame buffer, artificially inflating space.

  Or to look at it another way - People were running X back when 4 megs
was a huge amount of memory. It must be *possible* to cram X down into a
limited memory space, whether or not current implementations do so by
default.


  As for GTK... GTK would be largely useless for the type of game I'd
write, and gaming would be my only use for a graphics layer. GTK would be
a great framework for applications - but a game would be displaying one
big bitmap in a fullscreen window, and little else. Most of GTK's features
are redundant in this context.

  So for writing word processors for the Dreamcast, by all means; but
either X or (preferably) a dumb framebuffer layer would be more suitable
for, say, porting Quake.


> 	Other places for research would be audio, I haven't seen any
> 	references to diddling with the audio system on a Dreamcast.

  I skimmed the KOS pages, and they have sound support if I remember
correctly.


  This brings up another point that I mentioned previously - Why isn't
the NetBSD/Dreamcast project clean-rooming advances made by other
Dreamcast projects? This would greatly reduce the amount of effort needed
for development of features that have already been researched.

Ttyl,
				-Christopher Thomas