Path: utzoo!attcan!uunet!mcvax!steven
From: steven@cwi.nl (Steven Pemberton)
Newsgroups: comp.lang.misc
Subject: Re: Modern languages (Slight return)
Message-ID: <352@piring.cwi.nl>
Date: 8 Jun 88 12:33:56 GMT
References: <3292@enea.se> <2414@uvacs.CS.VIRGINIA.EDU> <11671@ut-sally.UUCP> <12753@shemp.CS.UCLA.EDU> <1090@micomvax.UUCP>
Reply-To: steven@cwi.nl (or try mcvax!steven.uucp)
Organization: CWI, Amsterdam
Lines: 41

In article <12753@shemp.CS.UCLA.EDU> gast@lanai.UUCP (David Gast) writes:
[In response to Ed Nather's reference to Tony Hoare]
>
> You might also want to remember that his reviews of Algol 68 were
> notorious for misreprenting the truth.  For example, he would give an
> example of something that was terribly wrong with Algol 68; the
> only trouble was that his terrible example would be a syntactic
> error caught by the compiler.
> 
In article <1090@micomvax.UUCP> ray@micomvax.UUCP (Ray Dunn) replies:
> Foul Play!
> 
> Does "he would give an example" actually mean "he gave (one) example" or, as
> it implies, that he *habitually* did this.

No indeed, David Gast is right. Hoare's criticisms of Algol 68 were
extremely misinformed and weren't confined to a single occurrence.
In fact 'foul play' is exactly what I thought when I read his reports
containing the misrepresentations (you must remember that at the time
there were language wars going on, mostly three-pronged
Pascal-Algol68-PL/I).

As an example, he reported that at a conference someone asked if Algol
68 supported triangular arrays, and the reply came that not only did
it support triangular arrays, but diamond shape, circular, or any
other shape you liked. Hoare then mocked this, by saying that he
didn't believe that there'd ever be an application that needed
circular arrays.

This annoyed me for two reasons. Firstly because I didn't see any
reason to reject the idea of circular arrays just on Hoare's say so,
but more importantly because of his suggestion that circular arrays
were a feature of Algol 68, rather than an outcome of a general, and
simple, mechanism (just as in Pascal, you can have arrays of arrays,
but unlike Pascal, the bounds of the arrays can be determined at
run-time, therefore two dimensional arrays can have any shape you want).

On the other hand, I found that many of Hoare's remarks on Ada hit the
mark.

Steven Pemberton, CWI, Amsterdam; steven@cwi.nl (mcvax!steven.uucp old style).