Received: from localhost (daemon@localhost) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id QAA07838; Sun, 30 Jun 1996 16:55:42 -0400 Received: by CS.UTK.EDU (bulk_mailer v1.6); Sun, 30 Jun 1996 16:55:09 -0400 Received: from koobera.math.uic.edu (qmailr@KOOBERA.MATH.UIC.EDU [128.248.178.247]) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id QAA07722; Sun, 30 Jun 1996 16:55:06 -0400 Received: (qmail-queue invoked by uid 666); 30 Jun 1996 20:58:35 -0000 Date: 30 Jun 1996 20:58:35 -0000 Message-ID: <19960630205835.29571.qmail@koobera.math.uic.edu> From: djb@koobera.math.uic.edu (D. J. Bernstein) To: drums@cs.utk.edu Subject: Re: Obsolete syntax > I support option 2 as well. Me too. I agree with Keith that there should be a grammar saying what you have to accept, and text saying what you can generate. I agree with Paul that it's a Bad Idea to have two conflicting grammars. Note that foo/obsolete-foo is only one side of the coin; the other side is foo/new-fangled-foo. An extreme example: year = obsolete-year / current-year / new-fangled-year obsolete-year = 1*3DIGIT current-year = 4DIGIT new-fangled-year = 5*DIGIT A parser must accept all of these, interpreting obsolete-year by the rules that we worked out. A generator must not create obsolete-year because it'll start confusing people in the year 2000. A generator must not create new-fangled-year because some current parsers can't handle it; but this rule is scheduled to change in the year 5000. ---Dan