From: usotsuki@buric.co   
      
   So here's what I really had in mind when I first started bringing up the   
   terrapin talk a couple weeks ago...   
      
   A friend of mine and I are designing a little pocket computer based on a   
   65C02 (I had the idea, he took it and ran with it; he's designing it at   
   the hardware level, but the firmware is mainly going to come from me).  We   
   were talking about how to handle the video mode, and since he and I had   
   already developed one of the two working emulators for the Nabu, we wound   
   up deciding to just clone the Nabu's graphics chip, since we both had a   
   fair idea how the thing worked.   
      
   Guess what... that makes it pretty similar to a very specific   
   configuration of an Apple ][!  Specifically, an Apple ][ with a sprite   
   card.  That means I can do some theoretical code testing while I don't   
   have a workable simulation of the system, using existing resources, such   
   as MAME.   
      
   So...   
      
   I had in mind to try to pack as much as I could into the ROM space, which   
   would be somewhat larger than that of the ][ (would be 32K, actually, and   
   unbanked).  And I was thinking, at first, the usual - you know, the kind   
   of stuff we have on the ][ line - monitor, miniassembler, BASIC, and we   
   started to think about what else might be viable, and Logo had been tossed   
   around as an idea.  That's why I was trying to figure out how to implement   
   turtle graphics - because ultimately I was going to have to implement   
   turtle graphics on a TMS9918/65C02 pairing.   
      
   That's also going to lead to another question...  Waitaminute, the TMS9918   
   doesn't have anything like the ]['s split screen graphics mode!  Actually,   
   that can be fudged, with a slight limitation.  G2 mode basically splits   
   the screen into three 256x64 windows, each implemented as a 32x8 set of   
   tiles.  32*8 = 256 - so often a person will just write 0-255 into each   
   set, then dink the font to create an all-pixels-addressable mode.  In this   
   case, that would only be done for the first 2 blocks, then 128-255 in the   
   top half of the third, and a standard ASCII font loaded as 0-127 of the   
   third tile set.  You lose the ability to do inverse text (there's still   
   ways to do this, by dinking the color memory), but that's still plenty   
   enough if you need graphics and text together.  You get 256x160 graphics,   
   plus 32x4 text.   
      
   So testing will probably be done initially on an emulated ][ - and that   
   will be enough to make sure the _logic_ works.   
      
   Ultimately, I was wanting to create an environment as similar as possible   
   to an Apple //e, within the limits of the hardware I was targetting - and   
   have it, you know, a monitor, mini-assembler, BASIC and Logo - and Logo   
   could avail itself of the BASIC in order to implement some of its   
   features, they could shere their floating point library and their graphics   
   support (which in Logo's case would be used to implement turtle graphics).   
      
   Once I brought in graphics and Logo, though, that's where my brain melted.   
   (I'll probably need to do preliminary implementations in some other   
   language and then translate it into 65C02 postfacto, but I know nothing of   
   low-level handling of graphics like Bresenham algorithm, circle   
   algorithms, etc., and might need a hand on these things.)   
      
   The Logo implementation, if it should come to fruition, would also be   
   ported to the Apple since there would be practically no further effort   
   required.   
      
   -uso.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)