Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!snorkelwacker.mit.edu!paperboy!macrakis
From: macrakis@gr.osf.org (Stavros Macrakis)
Newsgroups: comp.lang.c
Subject: Re: Request for CMS2-to-c converter
Message-ID: <MACRAKIS.91Jan30034156@diderot.gr.osf.org>
Date: 30 Jan 91 08:41:56 GMT
References: <62394@bbn.BBN.COM>
Sender: news@OSF.ORG
Organization: OSF Research Institute--Grenoble
Lines: 36
In-reply-to: dcousins@bbn.com's message of 29 Jan 91 19:05:43 GMT

In article <62394@bbn.BBN.COM> dcousins@bbn.com (Dave Cousins) writes:

   Request for CMS2-to-c converter

   Does anyone know of such a beast? A CMS2 to 68020 cross compiler or
   CMS2 to fortran could also work.

Some years ago, I studied CMS-2 and some sample CMS-2 programs in
order to understand the transition to more modern languages and
programming techniques, namely Ada.

In my considered opinion, it would be a big mistake to try to convert
CMS-2 programs to anything else without first redesigning.  CMS-2 is a
very low-level language, whose semantics depends very closely on the
particular compiler and target machine (and especially the word size).
The CMS-2 code I saw `took advantage' of this in many ways, thus
making the code very difficult to understand and maintain.  In
particular, type definitions in CMS-2 were typically used as synonyms
for memory layout, and assembly instructions were freely interspersed
with CMS-2 code.  Perhaps the code you have is cleaner.

There are also some specific features of CMS-2 which are hard to deal
with.  Take a look at the feature which lays out an array of records
as a record of arrays, for instance.

Even if I had a translator or cross-compiler for CMS-2, I wouldn't
trust the result.

	-s

Stavros Macrakis
Open Software Foundation Research Institute

Mail:  2 av de Vignate, 38610 Gieres (Grenoble), France
Net:   macrakis@gr.osf.org
Phone: +33/76.63.48.82  Fax: +33/76.51.05.32