X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil] ["4821" "Sat" "5" "December" "92" "19:57:32" "+0100" "Lutz Birkhahn" "lutz@BISUN.NBG.SUB.ORG" nil "87" "Re: default paper sizes" nil nil nil "12"]) Return-Path: Received: from sc.ZIB-Berlin.DE (serv01) by dagobert.ZIB-Berlin.DE (4.1/SMI-4.0/1.9.92 ) id AA17270; Sun, 6 Dec 92 06:22:16 +0100 Received: from vm.urz.Uni-Heidelberg.de (vm.hd-net.uni-heidelberg.de) by sc.ZIB-Berlin.DE (4.0/SMI-4.0-sc/19.6.92) id AA04380; Sun, 6 Dec 92 06:22:13 +0100 Message-Id: <9212060522.AA04380@sc.zib-berlin.dbp.de> Received: from DHDURZ1 by vm.urz.Uni-Heidelberg.de (IBM VM SMTP V2R2) with BSMTP id 2728; Sun, 06 Dec 92 06:22:37 CET Received: from DHDURZ1 by DHDURZ1 (Mailer R2.08 R208004) with BSMTP id 2322; Sun, 06 Dec 92 06:22:33 CET Received: from DHDURZ1 by DHDURZ1 (Mailer R2.08 R208004) with BSMTP id 2320; Sun, 06 Dec 92 06:22:29 CET Reply-To: Mailing list for the LaTeX3 project In-Reply-To: Paul Taylor's message of "Thu, 26 Nov 92 12:27:41 +0100." Date: Sat, 5 Dec 92 19:57:32 +0100 From: Lutz Birkhahn Sender: Mailing list for the LaTeX3 project To: Multiple Recipients of Subject: Re: default paper sizes Status: R X-Status: X-Keywords: X-UID: 899 > The hypothesis was that we have a CD rom with 15000 documents on it [in source]. > FTP archives give rise to a similar situation. In this case I don't care how > the documents are typeset - I just want to be able to see as quickly as possible > what's in the collection. Maybe later I'll want to print a few of them > "perfectly". OK, perhaps my view was too much that of a perfectionist. There are several different "perfection levels" of documents: at the one end of the range is the real work of art, where every dimension is carefully determined, every paragraph tuned to come out well, and so on. On the other side there is the "every-day work", e.g. technical reports or similar document types, where the main goal is to have a usable output that is at least readable and correct (in a technical sense), but not much more. Here it doesn't matter much if e.g. the paper size is "letter" or A4, if the inter- word space is large (very loose lines) or not, nor where the text is positioned on the paper (i.e. left and top margins). I can imagine several approaches to deal with these different "perfection levels". For example, there could be a declaration at the beginning of the document, stating whether this is a "piece of art" or not. Perhaps it could be more concrete, e.g. in the sense of "this document is fine tuned to be printed on A4 paper size, don't try to use different paper size, if ever possible. Give a severe warning to the user if he wants to use a different paper size (like e.g. the "over- full hboxes" messages), using some drastic words. But if the user really wants to, accept his own size specifications." The thousands of documents that are on your CD will probably not have this "high quality tag", so you can specify any paper size you want without having TeX to complain about every document. One sort of "low quality tag" in LaTeX2 is the \sloppy declaration. Unfortunately its effects are very bad: people often wonder why there are some lines with only two words in it and a huge amount of white space between them (a consequence of the infinite \tolerance). Clearly this is an error that should be corrected in LaTeX3, which should be no problem with the new \emergencystretch parameter of TeX 3.x. > Increasing the page width is in fact likely to reduce the number of > over-full \hbox'es. ... and likely to reduce legibility, by the way. At least in some cases. Only to remind you that page size is *not only* a question of economy or paper size. > So, I think the original suggestion, that the paper size should be settable by > some external method (such as looking for "paper-size.sty") is reasonable. Given the above, I agree that it could be fine in *some* circumstances to have the ability to specify some document parameters without changing the document itself. My main objection concerned Tarjei T. Jensen's proposal to solve this with a (operating system dependent!) specified order of different styles with the same name, but from different dircetories. Besides, I think LaTeX3 should work with today's TeX3, not (only) with NTS (New Typesetting System, sort of successor to TeX) or whatever it's name will be. And today's TeX3 does not read multiple styles with the same name, although some imple- mentations (but not all!!) allow you to specify more than one directory where to look for files. With today's TeX there might be some solutions: one mentioned by Paul Taylor is a "paper-size.sty" (for portability reasons, the name should be shorter and without a hyphen, if possible) or something like a "local.sty" or "local.tex". A second possibility could be TeX commands given in the command line, e.g. in Un*x latex '\papersize=a4 \input document' then the style could check that \papersize variable (and spit out a warning if needed). I'm not sure if this solution still meets the portability criterion. A third solution could be some actions in the \everyjob token list, e.g. loading some file, although I see no advantages over the first solution. I'd favor the second solution, if it proves to be portable. I dislike solu- tions that change the paper size automagically, since I think that document distortions at least should be specified *explicitly*. One of the main reasons that most DTP documents look so bad is that all bad things are as easy to do as the good ones (sometimes even easier). TeX and LaTeX should continue to select the best quality by default, even if used as a formatter of "every-day documents". -- Lutz Birkhahn (Germany) email: lutz@bisun.nbg.sub.org (don't use another!) F"urther Str. 6 +-------------------------------------------------- D-W-8501 Cadolzburg 2 | "It is an error to not have enough arguments" Voice: 09103 / 2886 | (Hype Programmer's Guide)