🚀 go-pugleaf

On 6 Nov 90 20:48:15 GMT, brnstnd@kramden.acf.nyu.edu (Dan Bernstein) said:

brnstnd> In article <6PX6-O@xds13.ferranti.com> peter@ficc.ferranti.com
brnstnd> (Peter da Silva) writes:

peter> What's wrong with:
peter> 	if(array_a_and_array_b_overlap)
peter> 		use_safe_code
peter> 	else
peter> 		use_fast_code_that_doesn't_work_for_aliased_arrays.

A lot of things. Code bloat for one. But it is an appealing solution to
some difficult problem in many useful cases.

brnstnd> [ ... ]
peter> The whole alias problem is a compiler technology problem.
brnstnd> Exactly.

No, and here we have our largest difference: it is a language design
problem, and a problem of choosing the suitable language.

This discussion reminds me of Dijkstra's famous address in 1968 to some
(the IFIP?)  conference, on the software crisis, that went like this:

  We have a software crisis, and we are told that we must advance the
  state of the art so as to solve it. We are only given a few constraints
  to satisfy:

	  no changes to the hardware platforms
	  no changes to the languages
	  no changes to the operating systems 
	  no changes to the education of programmers
	  no changes to the software tools
	  no changes to organizational practices
	  ...

  Well, this is RIDICULOUS!

In twenty two years progress has not exactly astonishing, except for a
few things, like DBMSes. There has even bit some kind of regress towards
mediocrity.
--
Piercarlo Grandi                   | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth        | UUCP: ...!mcsun!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk