From richard@rrbcurnow.freeuk.com Sun Jul 22 15:07:46 2001 Return-Path: X-Sender: richard@rrbcurnow.freeuk.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_2_0); 22 Jul 2001 22:07:45 -0000 Received: (qmail 73160 invoked from network); 22 Jul 2001 22:07:40 -0000 Received: from unknown (10.1.10.27) by l8.egroups.com with QMQP; 22 Jul 2001 22:07:40 -0000 Received: from unknown (HELO scrabble.freeuk.net) (212.126.144.6) by mta2 with SMTP; 22 Jul 2001 22:07:40 -0000 Received: from du-009-0146.freeuk.com ([212.126.152.146] helo=rrbcurnow.freeuk.com) by scrabble.freeuk.net with esmtp (Exim 3.12 #2) id 15ORNy-00012t-00 for lojban@yahoogroups.com; Sun, 22 Jul 2001 23:07:38 +0100 Received: from richard by rrbcurnow.freeuk.com with local (Exim 2.02 #2) id 15OQ4J-000021-00; Sun, 22 Jul 2001 21:43:15 +0100 Date: Sun, 22 Jul 2001 21:43:15 +0100 To: "Lojban@Yahoogroups. Com" Subject: Re: [lojban] the formal grammars' utility Message-ID: <20010722214315.B111@rrbcurnow.freeuk.com> Mail-Followup-To: "Lojban@Yahoogroups. Com" References: <4.3.2.7.2.20010714124253.00b6c2b0@127.0.0.1> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2i-nntp In-Reply-To: <4.3.2.7.2.20010714124253.00b6c2b0@127.0.0.1>; from lojbab@lojban.org on Sat, Jul 14, 2001 at 12:46:59PM -0400 From: Richard Curnow Apologies for following up a week late - I've been on vacation! On Sat, Jul 14, 2001 at 12:46:59PM -0400, Bob LeChevalier (lojbab) wrote: > At 04:40 PM 07/14/2001 +0100, And Rosta wrote: > > >3. Has anybody created a more succinct but unabbreviated (and, ideally, > >more intuitive) version of the YACC grammar? > > You just said why people use the EBNF. If the YACC grammar could be more > succinct, we would. The EBNF covers the whole language, so it is more > succinct and unabbreviated. Its weakness is that it is not proven > equivalent to the YACC, so mistakes in obscure places could still surface. > When I created the yacc/bison input for jbofi'e, I deliberately worked from the EBNF file rather than John Cowan's yacc input file. My hope was that I would have an independent yacc grammar, and in the long term this would help us find differences between the ENBF and official yacc version. During jbofihe's lifetime, remarkably few goofs have been found. Even then, the problems have tended to be more with the pre-parsing stage; the most recently found fault was that "na ku" has to be pre-parsed to appear as a single token to the yacc stage. The above is a rather hubristic view, of course, since most Lojban texts do not exploit the more esoteric parts of the grammar or combinations thereof at all. A project I really want to do is to 1. get the "official" v3.00 parser building on Linux 2. complete a random Lojban sentence generator I have which is 1/2 finished 3. write some scripting to feed random sentences to the official parser and jbofihe in parallel, to find cases where the parsers disagree on legality. (I'd begin with allegedly legal sentences and then find subtle ways of corrupting them.) Unfortunately, this job has been "pending" for quite some time now. -- R.P.Curnow,Weston-super-Mare,UK |lo samskiro'a cu simsa lo'e glefau http://www.rrbcurnow.freeuk.com/ |isa'e le xamgu cu tcetcexau ije richard@rrbcurnow.freeuk.com |le xlali cu xagmau lenu nomei