From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from faui45.informatik.uni-erlangen.de ([131.188.1.45]) by hawkwind.utcs.toronto.edu with SMTP id <2237>; Fri, 17 Sep 1993 13:33:04 -0400
Received: from faui30.informatik.uni-erlangen.de by uni-erlangen.de with SMTP;
	id AA16599 (5.65c-5/7.3v-FAU); Fri, 17 Sep 1993 19:32:50 +0200
Received: from faui30a.informatik.uni-erlangen.de by immd3.informatik.uni-erlangen.de with SMTP;
	id AA01259 (5.65c-5/7.3v-FAU); Fri, 17 Sep 1993 19:32:44 +0200
From:	Stefan Dalibor <Stefan.Dalibor@informatik.uni-erlangen.de>
Message-Id: <199309171732.AA01259@faui30.informatik.uni-erlangen.de>
Date:	Fri, 17 Sep 1993 13:32:47 -0400
To:	rc@hawkwind.utcs.toronto.edu
Subject: Re: list operators
Reply-To: Stefan.Dalibor@informatik.uni-erlangen.de


Hi,

as it seems to be list-op wishing-time, here are my favourites:

1. Cartesian products of lists; it can of course be done with
   functions (and I can live with that in scripts, though I don't like
   it), but I use this *very* often as globbing construct and having
   to type the rc equivalent of e.g.
    what {{$O/bin,~/bin}/{sun*,m88k,ptx},src/{zrc,xvt}}/{zrc,rc,sysconf,xvt}
   interactively can be extremely tedious.
   IMHO it seems also a bit inconsequent to implement cartesian
   products only for the case of 1 x n elements; I'd vote for
   implementing the general n x m case which includes the former.

2. Assignment to list elements, e.g. `path(5)=/local/bin'. I think
   there should be no need to use a function for such an atomic
   operation.

Both items can be justified by (and will of course be rejected for) the
same reasons as the earlier proposals; I posted them yet because much
to my surprise, the absence of these constructs (and not the usual
like job-control, user-name-expansion, completion, builtins etc.) was
the reason that one of the few colleagues I was able to interest in rc
dropped it finally and keeps on using tcsh... so I think this can at
least be filed as a curiosity.

Bye,
Stefan