Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!yale!cmcl2!lanl!jlg
From: jlg@lanl.gov (Jim Giles)
Newsgroups: comp.lang.misc
Subject: Re: Re: Aggressive optimization vs HLL's
Message-ID: <6074@lanl.gov>
Date: 15 Nov 90 22:38:11 GMT
References: <7226:Nov1521:52:4690@kramden.acf.nyu.edu>
Organization: Los Alamos Natl Lab, Los Alamos, N.M.
Lines: 50

From article <7226:Nov1521:52:4690@kramden.acf.nyu.edu>, by brnstnd@kramden.acf.nyu.edu (Dan Bernstein):
> In article <6054@lanl.gov> jlg@lanl.gov (Jim Giles) writes:
>> No, actually it doesn't.  Not if the language has a "whole array" syntax.
>> In such a language, you pass the _specific_ section of the array that
>> the next procedure needs.
> 
> Are you implying that Ada, which has this feature (under the somewhat
> misleading name of ``slicing''---it's a type of slicing, but a very
> restricted type), suffices for array shifts?

No.  I was responding to a specific example (the context of which
you have deleted).  I try to reflect the context of what _you_ say
as accurately as possible.  This tends to make you all the more
angry.  Take the case where you strongly attacked arrays (as opposed
to pointers) and in the _same_ article stated that high-level
structures should be preprocessed to pointers.  This prompted me to
make the assumption (correct according to context) that arrays were
among the things you wished to preprocess out.  You _now_ claim not
to have meant that.  But, I was _trying_ to respond to the whole
gist of your statement _including_ context.

Now, the _specific_ example I was responding to here did not refer to
the concept of array shifts but merely to the concept of precomputing
an array base + offset and using that as the _base_ of an array in
a different routine.  Period.  I was not implying that ADA could
answer the problem referred to in the specific example given.  I was
implying nothing more (not about ADA anyway - which I didn't mention
by name, or any other extant language for that matter).  In actual
fact, Fortran Extended was the model for the "whole array" syntax
that I referrred to.

> [...]
> How do you take an array x indexed from 0 to 10 and pass it to a
> procedure indexed from -5 to 5 (presumably with a pointer to the 0th)?
> I don't see a way to do this.

Again, the Fortran Extended mechanism can do this.

> [...]
> If it's there, my apologies; this is what I get for not spending a lot
> of time with a language before talking about it. In that case the word
> ``Ada'' should be deleted from my postings on arrays versus pointers.
> But I just don't see how it's done.

That's because there is only _one_ person on the net who refuses to
look at anything but _extant_ languages.  In _your_ thread, I'm
willing to conform to that rediculous constraint on the nature of
the discussion, but not here.

J. Giles