Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!pt.cs.cmu.edu!dravido.soar.cs.cmu.edu!acha
From: acha@CS.CMU.EDU (Anurag Acharya)
Newsgroups: comp.lang.functional
Subject: "off-side rule"
Message-ID: <ACHA.91Jan17120316@DRAVIDO.CS.CMU.EDU>
Date: 17 Jan 91 17:03:16 GMT
Sender: acha@dravido.soar.cs.cmu.edu
Distribution: comp
Organization: Carnegie Mellon University
Lines: 20


Thanks for all the info on the "off-side rule". As far as I could make out,
the rule is motivated by a desire to make the syntax of programming languages
look as close to that of mathematical functions as possible. As the post
on Haskell mentioned, the layout is preprocessed into a set of tokens
in a phase between the lexer and the parser.
Such functionality would be better placed in an editor rather than in the 
compiler. We already have the "electric-this", "electric-that" modes in 
gnu-emacs. It should be a simple job to extend these modes to include the 
"off-side rule" if so desired. Why complicate the language definition
with such inelegant constructs ?

Of course, if one is indeed going to depend on the programmers to do the 
indentation right, one might expect them to put in the delimitors. A point
that is made in the defence of the "off-side rule" is that the indentation,
if "properly" done, renders the delimitors redundant -- so we should do away 
with them. This is, however, useful redundancy -- in that it allows the
compiler to detect any inadvertent errors that might occur.

anurag