From bob@RATTLESNAKE.COM Mon Aug 27 16:36:23 2001 Return-Path: X-Sender: bob@rattlesnake.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_3_2); 27 Aug 2001 23:36:23 -0000 Received: (qmail 22386 invoked from network); 27 Aug 2001 23:33:44 -0000 Received: from unknown (10.1.10.26) by l8.egroups.com with QMQP; 27 Aug 2001 23:33:44 -0000 Received: from unknown (HELO localhost) (140.186.114.245) by mta1 with SMTP; 27 Aug 2001 23:33:43 -0000 Received: by rattlesnake.com via sendmail from stdin id (Debian Smail3.2.0.111) for lojban@yahoogroups.com; Mon, 27 Aug 2001 23:33:36 +0000 (UTC) Message-Id: Date: Mon, 27 Aug 2001 23:33:36 +0000 (UTC) To: jay.kominek@colorado.edu, lojban@yahoogroups.com Cc: bob@rattlesnake.com In-reply-to: (message from Jay Kominek on Mon, 27 Aug 2001 17:02:30 -0600 (MDT)) Subject: Re: [lojban] LALR1 question Reply-to: bob@rattlesnake.com References: From: "Robert J. Chassell" LR(k) and LL(k) (for all k) grammars are also unambiguous. Like Robert McIvor, I, too, was under the impression that 20 years ago people were not sure whether the LR(k) parsers were truly unambiguous. That is when the relevant grammar choices were made (1970s, early 1980s). The only tool that generated confidence was YACC, which had only recently been developed. If I remember rightly, it worked only with LALR(1) grammars. YACC could not handle the Loglan grammer directly, but required a preproccessor that might not have been unambiguous itself. I could well be wrong, but that is what I remember. Also, very practically, no one in the project had fast machines back then. -- Robert J. Chassell bob@rattlesnake.com Rattlesnake Enterprises http://www.rattlesnake.com