Re: Graphics Compilers?

sethml@ugcs.caltech.edu (Seth LaForge)
22 Apr 1997 21:26:43 -0400

          From comp.compilers

Related articles
Graphics Compilers? jukkaj@ping.at (JUKKA) (1997-04-18)
Re: Graphics Compilers? rweaver@ix.netcom.com (1997-04-20)
Re: Graphics Compilers? karsten@tdr.dk (Karsten Nyblad) (1997-04-20)
Re: Graphics Compilers? meyer@lps.u-psud.fr (peter meyer) (1997-04-20)
Re: Graphics Compilers? af137@torfree.net (1997-04-20)
Re: Graphics Compilers? sethml@ugcs.caltech.edu (1997-04-22)
Re: Graphics Compilers? d92-dne@nada.kth.se (Daniel Neri) (1997-04-22)
Re: Graphics Compilers? case@tvol.it.tvol.it (1997-05-04)
| List of all articles for this month |

From: sethml@ugcs.caltech.edu (Seth LaForge)
Newsgroups: comp.compilers
Date: 22 Apr 1997 21:26:43 -0400
Organization: California Institute of Technology, Pasadena
References: 97-04-119
Keywords: visual

On 18 Apr 1997 01:11:19 -0400, JUKKA <jukkaj@ping.at> wrote:
>Is there any compiler that takes some kind of a graphics
>representation as an input and produces from that the code?


UC Berkeley and Caltech have been working on this for years. At this
point we have a program, magic, in which you draw a symbolic
representation of your program using colored rectangles, and use a
program named COSMOS to compile your program to C. For additional
convenience, programs created in magic can easily be implemented
efficiently in silicon.


For those who haven't caught on yet, these are VLSI design tools. Not
very useful for ordinary programming, but terribly useful if you
happen to want to make a chip. You lay out your chip by drawing
doping and other stuff: polysilicon here, n-doping here, etc. You can
define cells, which are rectangular portions of layout, and compose
them to produce more complicated structures. Procedural programming
in VLSI!


I'm sure this isn't quite what Jukka had in mind, but it does seem to
fit his requirements quite nicely...


Seth
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.