home .. forth .. colorforth mail list archive ..

Re: [colorforth] colorforth.com pages, and other things.




Samuel Falvo wrote:

I visited ColorForth.com, and under the VLSI section, I note that a number of
links and images are broken.
Better late than never;p censorship always bugged me, perhaps the fact that the NSA got the tile definition is for something, they block patents, so maybe this stuff is too hot.
Too hot for what? you tell me.

Also, I haven't heard much of anything from this list in a long time.  And
certainly not from Chuck Moore and Jeff Fox.  Is it possible to arrange another
fireside chat?  If a video presentation isn't possible, perhaps a webpage can
be constructed in its place, containing the images and dialog in a more or less
slideshow format.
My dream is almost coming true.
Better, simpler threading and compilation scheme.
I have some printed code will show you soon.
(had horrible relative/absolute address bug in previous pushopt optimization lol)

: test bla1 bla2 bla3 ;  ->Old way
bla1bla2bla3test ->new way, with colors and i hope no spaces (same-color ambiguity but still)
                             test is red others are green and/or yellow
I eliminated ; and calls, now redwords compile pushes from-until nonzero item on data stack, convert the last one to jump to finish definition or compiles a bare ret opcode if theres
a zero on stack (no green words to compile)
would be if all green:
push3 push2 push1
push3 push2 jmp1  -> way better than "call1 call2 call3 ret"
*list of adress in dictionary could use this "reverse"scheme, who knows...

now for magenta words I must choose:
pushFORWARD push 3 push 2 jmp1 FORWARD: "magenta word bytes to copy here"
or
push all adresses on return stack, therefore inversing the order (321-123)and by doing a simple sub eax, edi I compile calls (origin of the major bug, call/jmp always relative to next instruction, push are absolute adresses and didnt do rel2abs conversion, so im taking
the other way around).
call1 call2 call3 smaller but maybe slower i dont know for sure.

Anyway i can explain better now its late and i got to get up at 5 am
hope i didnt confuse you too much, but my project is very exciting and new,
exactly as I wished.

Adam Marquis



---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com