Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!world!esegue!compilers-sender
From: hankd@ecn.purdue.edu (Hank Dietz)
Newsgroups: comp.compilers
Subject: Precise definition of "syntax diagram"
Keywords: parse, LL(1)
Message-ID: <9010180639.AA10335@dynamo.ecn.purdue.edu>
Date: 18 Oct 90 06:39:40 GMT
References: <9010091533.AA02386@apple.com> <9010101445.AA06181@dynamo.ecn.purdue.edu> <1990Oct16.015524.25858@comp.vuw.ac.nz>
Sender: compilers-sender@esegue.segue.boston.ma.us
Reply-To: hankd@ecn.purdue.edu (Hank Dietz)
Organization: Purdue University Engineering Computer Network
Lines: 30
Approved: compilers@esegue.segue.boston.ma.us

In article <1990Oct16.015524.25858@comp.vuw.ac.nz> lindsay@comp.vuw.ac.nz (Lindsay Groves) writes:
>In article <9010101445.AA06181@dynamo.ecn.purdue.edu>,
>hankd@dynamo.ecn.purdue.edu (Hank Dietz) writes:
>|> Pascal is generally defined using a set of syntax diagrams which, by
>|> definition, means the language can be recognized using LL(1).
>
>You must be using a strange definition a syntax diagrams if they can
>only describe languages that have LL(1) grammars.

The rule I have heard is that a syntax diagram is literally a flowchart
for an LL(1) parser where the branching decisions are obvious, hence
the branch decision nodes are omitted.  This was certainly true of all
syntax diagrams I saw earlier, but I have more recently seen "syntax
diagrams" which are no more constrained than arbitrary CFGs + looping,
and differ from them only in having arrows between symbols.

E.g., is the following a legal syntax diagram for S -> S a, S -> b?

S --+-->[S]---->(a)--+-->
    |                |
    +-->(b)--------->+

I believe it isn't.  Right or wrong?  References, anyone?

					-hankd@ecn.purdue.edu

PS: The unconstrained form looks more like a RTN to me....
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{ima | spdcc | world}!esegue.  Meta-mail to compilers-request@esegue.