Re: GNUstep and 256 colors.

[Gregory John Casamento]

Please see below...

--- Riccardo Mottola <multix@ngi.it> wrote:
> on 12/11/03 4:53 PM, Marko Riedel at markoriedelde@yahoo.de wrote:
> 
> > 
> > /*
> > * This block of code should be totally portable as it uses the
> > * 'official' X mechanism for converting from pixel values to
> > * RGB color values - on the downside, it's very slow.
> > */
> > 
> > 
> > You might have to make a case that there still are enough pseudocolor
> > visuals out there to make fixing this worthwhile.
> 
> There are! I didn't remember reading this thread, howewer, at least, I
> understand the problem better right now.
> Almost all unix workstation I had to mess up with that are some years old
> have 256 color cards. Maybe 2d accelerated, but 8bit (sun GX/TGX is maybe
> the most popular, but hp's pa9000 follow the same path)
> 
> Alsom the problem skyrockets when exporting to a 256 color display while
> locally, on a fast computer things are almost unusable.

Realistically, how many systems are using 256 color displays given today's
hardware?   The systems you are talking about must not just be a couple of
years old, but VERY old.

While I realize that it's important have as broad a support base as possible, I
haven't seen a 256 color display since I was using OPENSTEP/Mach on a 100MHz
Pentium (no blasted II, III, or IV).

> Thanks to Adam's patch which makes images appear on 256 colors, GNUstep is
> now almost usable on those older boxes when equipped with enough ram . Maybe
> asking for dithering would be too much from what I grasped of the look-up
> mechanism.

I imagine dithering would be difficult given the color model in X.  DPS used to
dither since it rendering was based on an internal color model which was
constant and would be transformed to be displayed on the output device (much in
the way your black and white printer simulates color shown on your screen when
you print).  DPS *always* rendered things internally in full color, but would
display them according to the capabilities of the device.

So far as I know this type of output device abstraction isn't present in X.

Later, GJC

=====
Gregory John Casamento -- CEO/President Open Logic Corp.
-- bheron on #gnustep, #linuxstep, & #gormtalk ---------------- 
Please sign the petition against software patents at: 
http://www.petitiononline.com/pasp01/petition.html 
-- Main Developer of Gorm (featured in April Linux Journal) ---

__________________________________
Do you Yahoo!?
New Yahoo! Photos - easier uploading and sharing.
http://photos.yahoo.com/