Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: $Revision: 1.6.2.13 $; site iuvax.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!ihnp4!inuxc!iuvax!willc
From: willc@iuvax.UUCP
Newsgroups: net.micro.apple
Subject: Re: C compiler
Message-ID: <1200002@iuvax.UUCP>
Date: Tue, 8-May-84 19:30:00 EDT
Article-I.D.: iuvax.1200002
Posted: Tue May  8 19:30:00 1984
Date-Received: Wed, 9-May-84 23:42:50 EDT
References: <1207@sdccs7.UUCP>
Lines: 22
Nf-ID: #R:sdccs7:-120700:iuvax:1200002:000:975
Nf-From: iuvax!willc    May  8 18:30:00 1984


I assume you're worried about the size of the parser, since any parsing
technique suitable for mainframes ought to work also on micros if the
program will fit.

If you have access to a parser generator such as yacc, use it and don't
worry about what kind of parser it generates.

Yacc happens to generate an LALR(1) parser.  A book on my shelf claims
that an LALR(1) parser table for Algol 60 occupies less than 3000 bytes.
I would guess that an LL(1) parser table would be about the same size.

A true recursive descent parser is just an LL(1) parser in which the
parser table is encoded procedurally.  I would expect table-driven LL(1)
parsers to be more compact than recursive descent parsers.

If you don't have access to a parser generator, use table-driven LL(1)
or recursive descent.  Just be careful as you transform the grammar into
LL(1) form.  Consult a good textbook on compiler construction such as
Aho & Ullman or Hunter.
				William Clinger
				...iuvax!willc