[GEM Development] New 256 colour driver

[John Elliott]

: 3) I'm somewhat surprised, myself, that the first -
: and second! (thanks, John) - 256 color drivers were
: for 1024x768 and *not* 640x480 -- till I read John's
: README file.

  My explanation only applies to my own SD256, not the 256-colour 8514/a
driver. For more on that, see below.

: (This has to do with the way the bytes 'line up', and
: are accessed, in video memory 'space' at one rez
: versus another. This, in turn, is related to real mode
: ONLY being able to access a 64KB (not more than 128KB)
: "window" of video memeory at a time. Addresses A0 00
: to AF FF or, up to, BF FF (Hex-a-decimal numbering).

  Some VESA BIOSes (not DOSEMU, which was what I was developing on) allow 
two 64k windows into video RAM, one at A0000 and one at B0000. 

  Now, 8514/a is a different kettle of fish. Unlike just about every other
video card on the planet, it doesn't appear in memory at all. Everything's
done by sending drawing instructions to I/O ports (though I'm not clear if
that's also the case for bit-blitting, or whether the card can directly
access motherboard RAM for that). Also unlike everything else, it really
does have eight planes rather than packed-pixel memory.  So there's no
technical reason why the Paradise driver couldn't have supported 640x480. I
suspect that the reasoning was that someone who bought an 8514/a clone
bought it because they wanted 1024x768x256 resolution, so that's what the
driver did.

-- 
John Elliott