Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!ucsd!rutgers!okstate!norman
From: norman@a.cs.okstate.edu (Norman Graham)
Newsgroups: comp.lang.misc
Subject: Re: Language Design
Message-ID: <4526@okstate.UUCP>
Date: 24 Mar 89 22:58:42 GMT
References: <9861@megaron.arizona.edu>
Organization: Oklahoma State Univ., Stillwater
Lines: 25

From article <9861@megaron.arizona.edu>, by gudeman@arizona.edu (David Gudeman):
> [stuff deleted about how the grammar specifies that
>  any identifier can serve as a type-specifier]
> 
> So both claims made in this argument are correct:
> 
> (1) the grammar of C is not context free
> and
> (2) the grammar given in K&R is context free

This is an interesting point (ie. one I hadn't thought of before :-).
In theory, a grammar is a generator for strings in a language, not
a recognizer for a language (eg. various forms of automata).  So
the grammar in K&R is context-free, but it doesn't generate the C
language--in addition to valid C programs it also generates invalid
C programs. Of course, when this grammar is used as the basis of
a compiler, the semantic analyzer should detect these invalid
programs. This makes sense to me... how about anyone else out there?


-- 
Norman Graham                            Oklahoma State University
  Internet:  norman@a.cs.okstate.edu     Computing and Information Sciences
      UUCP:  {cbosgd, ihnp4,             219 Mathematical Sciences Building
              rutgers}!okstate!norman    Stillwater, OK  74078-0599