Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!uunet!mcvax!hp4nl!ruuinf!kramer
From: kramer@ruuinf.cs.ruu.nl (mark kramer)
Newsgroups: comp.lang.misc
Subject: Re: syntax and semantics
Summary: The compiler doesn't need to "know" anything about semantics
Keywords: parse automated syntax semantic
Message-ID: <1257@ruuinf.cs.ruu.nl>
Date: 13 Apr 89 14:05:06 GMT
References: <2700@cuuxb.ATT.COM>
Organization: Univ of Utrecht, Dept of CS
Lines: 39

In article <2700@cuuxb.ATT.COM>, mmengel@cuuxb.ATT.COM (Marc W. Mengel) writes:
| The reason this discussion has generated so much discussion
| is that we have a clash between the theory-of-computation
| folks and the pragmatic programmer types. ....
| 
| The theoretical folks are talking about Context Free Grammars
Surely, CFG's are not powerful enough for programming languages, but
there are other types of grammars (type 0 or unrestricted grammars are as
powerful as any conceivable computer program; Church's hypothesis for
Turing Machines).

| To the theoretical folks ....            .... Semantics to them
| is the meaning associatecd with the text, i.e. "i = j;" has the
| effect of changing the value of the storage associated with the
| variable "i".
| 
| The pragmatic programmers are referring to syntax and semantics
| based upon the tools they use to build compilers.  ....
| ....  If our parser generators were powerfull enough
| to really parse the languages involved, this definition would
| coincide fairly well with the theoretical one; the code generated
| for a given sequence of characters expresses the semantics of that
| sequence.
No, the code generated is just another string of symbols.
The semantics is given by the execution on a target machine.
Thus the definition above 'coincides fairly well' with the theoretical
definition of *syntax*; code generation may be regarded as a purely
*syntactical* transformation, so (substitute transform for check):
|                                             .... we get code
| that is really checking syntax, but that our pragmatic programmers
| have fallen into the habit of calling "semantic code", because
| it wasn't built by the parser generator.

A member(element) of {theoretic folks} INTERSECT {pragmatic programmers}

Mark R. Kramer, Department of Computer Science, University of Utrecht
Padualaan 14,   P.O. Box 80.089,   3508 TB  Utrecht,  The Netherlands
Phone: +31 - 30 - 531937        |  UUCP    : ...!hp4nl!ruuinf!kramer
                                |  INTERNET: kramer@cs.ruu.nl