Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!unisoft!hoptoad!dasys1!aj-mberg
From: aj-mberg@dasys1.UUCP (Micha Berger)
Newsgroups: comp.lang.misc
Subject: Re: Language Design, or: is C's Grammar Context Free
Summary: Sol'ns
Keywords: BNF, Grammars, Silliness
Message-ID: <9313@dasys1.UUCP>
Date: 11 Apr 89 23:22:02 GMT
References: <5200040@m.cs.uiuc.edu> <12443@watdragon.waterloo.edu> <920@m10ux.ATT.COM>
Organization: The Big Electric Cat
Lines: 32

> Consider the following in C, however:

>	. . . { T (x); . . . }
 
> How should the construct be parsed?

It would have to be an expression, since anything within brackets ain't a
decleration. This isn't Pascal kiddees, all functions are global.

> Consider instead the one that someone else posted:
> 
> 	{ T * x; . . . }

This just shows that your grammar is poor. Obviously it has to be the
interpreter's job to distinguish between decleration and expression.

I don't see any qualitative difference between typedef's and variable
decleration. An identifier is an identifier. Are you going to tell me
that Pascal isn't CF because A*B could have diffent meanings, because
multiplying integers creates different code than multiplynig reals?

Another thing. No language is a CFG. A language just isn't a grammar.
In order to prove that C isn't A CFL, you'ld have to prove that there
doesn't exist a CFG that only accepts valid C programs. A grammar
doesn't care what you do with each statement, so who cares how it would
be parsed? You'll just get a very complicated interpreter, since very
little could be said about statement types in advance.
-- 
					Micha Berger				     
Disclaimer: All opinions expressed here are my own. The spelling, noone's.
email: ...!cmcl2!phri!dasys1!aj-mberg	       Aspaklaria Publications
  vox: (718) 380-7572			       73-32 173 St, Hillcrest, NY 11366