Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!uunet!mcvax!hp4nl!kunivv1!phoibos!ge
From: ge@phoibos.UUCP (Ge Weijers)
Newsgroups: comp.lang.misc
Subject: Re: Language Design
Message-ID: <540@phoibos.UUCP>
Date: 23 Mar 89 12:12:59 GMT
References: <9122@claris.com>
Organization: University of Nijmegen, The Netherlands.
Lines: 31

From article <9122@claris.com>, by hearn@claris.com (Bob Hearn):
>>I'm looking in the back of K&R (first edition), and if I'm not mistaken I'm
>>seeing context-sensitive stuff in the grammar.
> Enough of this bullshit!  How many times do I have to say it???  The grammar
> in K&R is context-free.  Do you know what a context-free grammar *is*?

Yeah, here is my regular C grammar (a regular expression, really):

c	:	
	|	letter c
	|	digit c
	|	/* etc. */
	;

Okay, it matches a lot more than C, but all legal C programs are acceptable.
So are all programs in all other programming languages. Useful!
The grammar in K&R is in the same category. It accepts all legal C programs,
plus a lot more. Because of current compiler technology we use an artificial
separation between syntax and static semantics. Static semantics can be
expressed in more powerful grammar types (2-level van Wijngaarden grammars etc.)
Your note is correct, but so is the observation that some context-sensitive
features are suggested by using different names for the same terminal, nl.
identifier. K&R use non-terminal names judiciously to suggest more than they
are really telling.
C can not be described exactly by a CFG, so C is not a context-free language.
Usually this type of confusion starts because one party is talking about
a grammar, whilst the other party is really talking about (an exact grammar of)
the language.

Ge' Weijers, KUN Nijmegen, the Netherlands
ge@cs.kun.nl