dwa@sdcc12.UUCP (Don Anderson) (09/02/85)
We are in the process of installing an NSF-funded national supercomputer center here at UCSD. The first processor will be a Cray XMP-48 (4 processors, 48 Megawords of memory, each word being 64 bits wide). We hope to add new machines as funding and the state of technology allow. The initial complement of software comes from the Labs (LLL and LANL) and it is heavily FORTRAN oriented, though there is a version of Pascal, a version of LISP and a version of C (the latter rather buggy and limited). For my own personal reasons - so that I could use it for a course that I teach - I'd like to have a Modula-2 compiler on the XMP. Is there anyone out there in netland who has some experience with installing M2 compilers who would be interested in taking this on? The XMP is, in many (but not all) ways a direct descendent of the old CDC 6600 for which the early ETH compiler was written, so it might be possible to adapt that. The San Diego Supercomputer Center will be bringing on remote sites in the first half of 1986, including Berkeley and Stanford with 56KBit/sec service. SDSC does not pay for the development of software, but it has considerable flexibility about the rules governing such development and does not require that the software be placed in the public domain in some circumstances. Anyone interested should contact me for particulars. -don anderson -director of academic computing, UCSD 619-452-2628 OAC B-028 UCSD, La Jolla, CA 92093
jrc@hpcnof.UUCP (09/16/85)
While I was going to grad school, I worked at the Idaho National Engineering Laboratory where, buried under a pile of papers and listings, was a Modula-2 compiler written in Pascal for a Cyber-like machine. The person who had the compiler is gone, but I recall it came from ETH. Jim Conrad hplabs!hpcnof!j_conrad
prl@ethz.UUCP (Peter Lamb) (09/23/85)
Modula 2 was indeed developed at ETH (Swiss Federal Institute of Technology)
by Prof. Wirth.
I have established a mailing address for people who want to get into direct contact
with the developers of this language.
{siesmo|decvax}!mcvax!cernvax!ethz!modula2
I hope that this will be of some assistance to people with questions about
the philosophy and definition of the language. I rely on the judgement of
the net users not to swamp the person who has taken this task on.
In response to the article from Kaare Christian regarding differences
between the definition of M2 between Ford&Wiener & Wirth, I have the following
from Wirth's group:
> Re: New Modula-2 Features mentioned in Ford&Wiener
>
> Taking Wirth's newest Modula-2 compiler (written by himself) as a reference
> I can give the following answers:
>
> 1) The unary minus is not allowed with sets. Use symmetric set difference
> instead. Example: {}/{1, 2}.
>
> 2) The syntax for the new if and while statements is:
> IfStatement = IF Expression THEN StatementSequence
> {(ELSIF | "|") Expression THEN StatementSequence}
> [ELSE StatementSequence] END.
> WhileSatement = WHILE Expression DO StatementSequence
> {"|" Expression DO StatementSequence} END.
> I don't know whether these changes are only of experimental nature
> or whether they will go into the language.
>
> 3) Structured function return values are allowed by the language but
> may be restricted in some implementations. Our compiler allows
> structured return values to the maximum size of 4 bytes. Be aware
> that no further selectings of structured return values are allowed
> by the Modula-2 syntax. Example: f(3)[7] is not allowed.
>
> 4) There's no change in syntax regarding the import. As an implementor
> I also see no benefits in changing the syntax in the proposed way.
>
> The following change in Modula-2 is considered in order to allow one pass
> compilation:
>
> ProcedureDeclaration = ProcedureHeading ";" (Block Ident | FORWARD).
>
>
> Werner Heiz
> Institut fuer Informatik
> Clausiusstr. 55
> CH-8092 ZUERICH
> Phone: 01 / 256 22 28
Peter Lamb ({siesmo|decvax}!mcvax!cernvax!ethz)
Institut fuer Informatik
CH-8092 Zuerich
Switzerland