Hmmm, your hardware may be a bit sparse, but you won't be taking up a
lot of cycles servicing the display either. If your processor was a 32
bit design you'd be golden - the Squeak object memory/gc design takes
good advantage of its 32 bit design and will be the hardest part to
change. On the plus side, very little in the Squeak image really cares
what is under it.

Is the memory addressing flat or segmented? Is the processor 32 bit
internally with a 16 bit external bus or 16 bit throughout?

Along with the PocketSmalltalk info, you might be interested in the
Implementation section of "Smalltalk 80: The Language and Its
Implementation" -- which I will be posting to my site this weekend (yes
really -- though probably without most of the figures to begin with; I
want to let people check it over for a bit as I'm finishing up before I
"officially" announce it; this was all supposed to be finished several
weeks ago, but anyway). This details the 16 bit implementation Squeak
was derived from (and was based on in its previous incarnation).

-- Dwight


Karsten Aalders wrote:
> 
> Hi,
> I am working on a PDA project.
> I am looking for a OS to use and Squeak may be a solution for that.
> The platform it should run on contains a 16bit/16MHz CPU (4mips), 2MB
> FPROM,
> 8MB RAM and a 320x200x1 B/W display (no gray-scale).
> The OS I use is WinNT/95. I have a ANSI-C compiler and a assembler for
> the platform.
> Can you give me some hints, if it is a good idea to portate squeak to
> my platform?
> I am not sure if there is enough performance, memory and man-power.
> 
> thanks in advance
> Karsten