X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1878" "Tue" "10" "June" "1997" "18:38:54" "-0400" "Michael John Downes" "mjd@MATH.AMS.ORG" nil "43" "Re: Multiple pass LaTeX3" "^Date:" nil nil "6" nil nil nil nil nil] nil) Received: from listserv.gmd.de (listserv.gmd.de [192.88.97.1]) by mail.Uni-Mainz.DE (8.8.5/8.8.4) with ESMTP id IAA10596; Wed, 11 Jun 1997 08:19:35 +0200 (MET DST) Received: from lsv1.listserv.gmd.de by listserv.gmd.de (LSMTP for OpenVMS v1.1a) with SMTP id <13.A814716A@listserv.gmd.de>; Wed, 11 Jun 1997 8:19:33 +0200 Received: from RELAY.URZ.UNI-HEIDELBERG.DE by RELAY.URZ.UNI-HEIDELBERG.DE (LISTSERV-TCP/IP release 1.8b) with spool id 151285 for LATEX-L@RELAY.URZ.UNI-HEIDELBERG.DE; Wed, 11 Jun 1997 08:19:29 +0200 Received: from math.ams.org (math.ams.org [130.44.210.14]) by relay.urz.uni-heidelberg.de (8.7.6/8.7.4) with SMTP id IAA02834 for ; Wed, 11 Jun 1997 08:18:12 +0200 (MET DST) Received: from epsilon.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IJX1HEVA5S0020FD@AXP14.AMS.ORG> for LATEX-L@RELAY.urz.uni-heidelberg.de; Tue, 10 Jun 1997 18:38:56 EST Received: by epsilon.ams.org; (5.65v3.2/1.1.8.2/12Oct95-1155AM) id AA31840; Tue, 10 Jun 1997 18:38:55 -0400 X-Mailer: Gnus v5.3/Emacs 19.34 Lines: 43 References: <199612231910.UAA16786@frank.zdv.uni-mainz.de> Message-ID: Reply-To: Mailing list for the LaTeX3 project In-Reply-To: Frank Mittelbach's message of Mon, 23 Dec 1996 20:10:35 +0100 Date: Tue, 10 Jun 1997 18:38:54 -0400 From: Michael John Downes Sender: Mailing list for the LaTeX3 project To: Multiple recipients of list LATEX-L Subject: Re: Multiple pass LaTeX3 Status: R X-Status: X-Keywords: X-UID: 2015 Hans Aberg wrote: > > For example, if one wants to automate the LaTeX standard two-pass > > compile, the environment "document" data would be put in a special file, > > which is read twice; if this feature is chosen, the first pass could then > > be done without actual typesetting, making it faster. To which Frank Mittelbach replied: > i fear that i don't understand what you mean. what "data" can be put > into a special file? most of the reason why you need the typesetting > several times (and it might be even more than two) is to gather > information from the visual context of the document, eg \pageref can > only report the correct number if everything up to that point is > already correctly typeset. I would like to see Hans's suggestion explored a little more, in the following directions: 1. If I never use \pageref in my document, the speedup of `syntax-only' typesetting on the first pass might be very welcome. (I.e., putting the entire document contents into the separate file.) 2. If all my cross-references are backward references, I would like to get my final dvi file after the first pass! 3. If the bibliography data is processed at the very beginning of a document in a sort of `numbers-only mode', then all \cite commands also could be back-references and I could get my final dvi file in one pass. 4. The texinfo idea of processing the table of contents at the end of the document would be nice (let the reordering of pages happen at the DVI driver level, if necessary; for that matter there is a sensible Russian convention to put the TOC at the end of a book anyway). Then maybe even the TOC, with correct page numbers, can also be gotten in one pass!! (Add exclamation points according to taste :-) (Hmm, I think I see a low-tech way to make a lot of this work even more simply, come to think of it ... ) Michael Downes, mjd@ams.org