Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.6.2.16 $; site ima.UUCP
Path: utzoo!watmath!clyde!cbosgd!ulysses!allegra!mit-eddie!think!ima!compilers
From: compilers@ima.UUCP
Newsgroups: mod.compilers
Subject: Some compiler questions
Message-ID: <136300070@ima.UUCP>
Date: Mon, 17-Feb-86 21:41:00 EST
Article-I.D.: ima.136300070
Posted: Mon Feb 17 21:41:00 1986
Date-Received: Wed, 19-Feb-86 04:07:53 EST
Lines: 25
Approved: compilers@ima.uucp
Nf-ID: #N:ima:136300070:000:1195
Nf-From: ima!compilers    Feb 17 21:41:00 1986


[from ihnp4!tektronix!reed!bart (Bart Massey)]

OK, here's a couple of compiler questions:

A)  Is 'recursive descent' really the evil parsing method some books
would have one believe?  If so, why? -- things I've read don't make it
very clear...

B)  Anybody know of grammar transformers or finite automata generators
running under UN*X?  I don't mean lex and yacc here -- I want a symbolic
representation of a MDFA or a BNF grammar as output, rather than a bunch
of @$%^ C code that I then have to go back and strip the logic out of.
The -v mode of yacc isn't TOO bad, but...

				Bart Massey
				..tektronix!reed!bart
[I always thought that recursive descent was a perfectly reasonable way to
build a compiler, particularly if you don't have an automatic parser generator.
A case can be made that they are better for error correction than bottom-up
parsers, since in a simple case you have a better idea of what you're looking
for.  Being able to feed a grammar to a generator program and having it hand
you back a parser that you know will work is awfully nice and I'd never go
back to manual parser writing if it didn't have to.  Responsible opposing
opinions invited, as they say.  -John]