Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!quintus!ok
From: ok@quintus.UUCP (Richard A. O'Keefe)
Newsgroups: comp.lang.c
Subject: Re: Refined C
Message-ID: <698@cresswell.quintus.UUCP>
Date: 26 Feb 88 12:47:05 GMT
References: <7619@pur-ee.UUCP>
Distribution: na
Organization: Quintus Computer Systems, Mountain View, CA
Lines: 30

In article <7619@pur-ee.UUCP>, hankd@pur-ee.UUCP (Hank Dietz) writes:
> 
> Several folk have requested that I post a brief  description
> of refined C...  note that conformant arrays are a subset of
> paramtypes.
> 2.  Status
> In addition,  we  have  a  prototype  tool
> called  CP  which,  using  rather complex and slow analysis,
> converts ordinary C code into reasonable (but not great)  RC
> code; CP is in the public domain.

Great stuff!  How do I get it?  {I've glanced at ICPP proceedings,
and am feeling uncommonly stupid for not having noticed this stuff.}
How do I get the relevant Purdue reports?

Re access rights: anyone here remember MARY (Mark(?) Rain's
Algol-68-like systems language.)  Access rights were part of types
in MARY, an unjustly neglected language.

Re conformant array parameters, however, no, conformant arrays are
not a subset of paramtypes.  When I proposed them for C, I was
already aware of parameterised types in ADA, Fortran 8X, and several
other languages.  I'll save my arguments against having the size of
an array being regarded as part of its type for another time (but
C doesn't do this for functions with integer arguments, why should
arrays be any more different from functions than they have to be?).
I suggested conformant array parameters rather than parameterised
types because I was trying very hard to avoid parameterised types!
(I think types with *type* parameters are wonderful, but that's
another story.)