Newsgroups: comp.sys.transputer
From: mveraart@fel.tno.nl (Mario Veraart)
Subject: Re: OCCAM-Compiler
Organization: TNO Physics and Electronics Laboratory
Date: Wed, 2 Mar 94 08:06:37 GMT
Message-ID: <1994Mar2.080637.15484@fel.tno.nl>

barry@cs.keele.ac.uk (B.M. Cook) writes:

>We're working on it.  We're getting there but we're not there yet.

>I know of 4 (potential) new routes to Occam on other than Transputer -

>1) The University of Twente has a learning package for their equivalent of
>  the British "Open University" which includes an Occam interpreter for
>  PC's.  There's a (small) cost involved.  Off the top of my head I can't
>  remember who to contact.

I have studied at University Twente and have worked with an OCCAM 2 
interpreter. It was one that was made in India if I remember well.
If this is the one they have today in the learning package then
I can say it is NOT workable. If you run a small program that
needs keyboard input then you have to tell the interpreter
many many times that there is still no key pressed. This requires
a keypress every time. It was very anoiing to work with.

The persons you can contact at University of Twente are
  prof. A. Bakkers or ir. J. Sunter at the Electrical engineering
faculty.

>2) Michael Poole has a working (but not final release) re-target to PC's
>  which produces '386 binaries (or '486 but not earlier versions since it
>  assumes a flat address space).  This uses the INMOS front-end.  At present
>  he is offering a cheap site licence for the current version.

>  Michael is not (yet) on email - I'd be happy to collect requests for
>  information and snail mail them to him.  I am "barry@cs.keele.ac.uk".

>3) SPOC - the Southampton Portable Occam Compiler.  This has been released
>  to testers and should be made public soon.  It converts Occam2 into C and
>  provides a micro kernel to run the code.  Currently available for Sun4's
>  (I don't know of any plans for a PC version).

>4) My own work on translating Occam3 into C is progressing well but is
>  generating questions about the definition of Occam3 - it will be a while
>  before these are resolved.  Benchmarks (i.e. read this with the usual
>  large pinch of salt) show this to be the fastest implementation at present
>  - a Sun4 goes more than twice as fast as a T8 (at many times the cost!).
>  Successfully runs on Sun's and PC's (and I'm looking seriously at
>  implementations for micro-controllers).

>Unfortunately funding is a problem for 2,3,4 to be developed further at any
>speed, they'll probably only continue slowly unless support can be found.

>I've not given a contact for 3) as I've not asked them if I may - I believe
>the distribution will be indirectly via an ftp site - watch this space.

>	    Barry M Cook.

Maybe I can add a fith route. I myself has written an OCCAM compiler
for the PC. It translates OCCAM to C that can be compiled with a
Turbo C compiler. The current state is as such that if I incorporate
strings and reals it can be used as a real working version.
Because of the time I still haven't done this. But someday I will.
Then it will be put in the public domain.

Mario
-- 
|Mario Veraart                  TNO Physics and Electronics Laboratory
|email: rioj7@fel.tno.nl        The Hague       The Netherlands
|"If all else fails, show pretty pictures and animated videos, 
| and don't talk about performance",  David Bailey

