Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!haven!adm!cmcl2!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.lang.misc
Subject: Re: Answers, Chapter 1: TeX
Message-ID: <4675@lanl.gov>
Date: 1 Nov 90 19:43:34 GMT
References: <4181@goanna.cs.rmit.oz.au>
Organization: Los Alamos Natl Lab, Los Alamos, N.M.
Lines: 38

From article <4181@goanna.cs.rmit.oz.au>, by ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe):
> In article <4569@lanl.gov>, jlg@lanl.gov (Jim Giles) writes:
>> [...]       An array is a mapping from one or more bounded index sets
>> to values of an underlying type.  Each array is assumed distinct from
>> all others unless _explicitly_ declared otherwise.
> 
> A consequence of this definition is that array cross-sections (as in
> PL/I, Algol 68, and Fortran Extended) are *NOT* arrays. [...]

Well, Fortran Extended array cross-sections certainly are not arrays.
If they are passed to subroutines, they can become arrays there.  The
reason they aren't arrays has nothing to do with aliasing though -
cross sections are not indexible (so they don't meet the frist part of
my definition of being mappings from index sets to values).

Fortran Extended pointers are aliased arrays (which can be aliased to
anything in their scope that has the target attribute and is the same
underlying type).  But, this constitutes an explicit declaration
(thought not explicit enough to my taste - I have opposed the new
standard) so they still fit into my definition (if only barely).

An array section passed to a subroutine must not be aliased to any
other argument or global variable that that is visible in the scope of
the subroutine (unless explicitly declared as usual).  In this case,
the section is considered an array again - it can be indexed, etc..
Once again, this fits my definition.

Still, you make a good point.  The operations allowed on array
sections should be carefully considered.  The purpose of making the
definition of arrays as I did was to allow the compiler (with very
little assist from the loader) to produce efficient code for array
manipulation and be assured that the optimizations are safe from
aliasing.  If some operations on array sections are possible which
preserve this property but are in contradiction to my definition of
arrays, then I am willing to add array sections as a separate data
construct to my list and define those new operations on them.

J. Giles