X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["3331" "Mon" "13" "December" "93" "12:17:36" "MST" "Grant Gustafson" "gustafso@MATH.UTAH.EDU" nil "72" "Re: ideas for floats (forwarded)" "^Date:" nil nil "12"]) Return-Path: Received: from sc.ZIB-Berlin.DE (mailserv) by dagobert.ZIB-Berlin.DE (4.1/SMI-4.0/24.6.93) id AA25209; Mon, 13 Dec 93 22:47:49 +0100 Received: from mail.cs.tu-berlin.de by sc.ZIB-Berlin.DE (4.1/SMI-4.0-sc/03.06.93) id AA10405; Mon, 13 Dec 93 22:47:39 +0100 Received: from tubvm.cs.tu-berlin.de by mail.cs.tu-berlin.de with SMTP id AA05889 (5.65c8/IDA-1.4.4(mail.m4[1.12]) for <@MAIL.CS.TU-BERLIN.DE:Schoepf@SC.ZIB-BERLIN.DE>); Mon, 13 Dec 1993 22:47:26 +0100 Message-Id: <199312132147.AA05889@mail.cs.tu-berlin.de> Received: from TUBVM.CS.TU-BERLIN.DE by tubvm.cs.tu-berlin.de (IBM VM SMTP V2R2) with BSMTP id 1048; Mon, 13 Dec 93 22:48:02 +0200 Received: from VM.URZ.UNI-HEIDELBERG.DE (NJE origin MAILER@DHDURZ1) by TUBVM.CS.TU-BERLIN.DE (LMail V1.2a/1.8a) with BSMTP id 1020; Mon, 13 Dec 1993 22:48:02 +0200 Received: from DHDURZ1 (NJE origin LISTSERV@DHDURZ1) by VM.URZ.UNI-HEIDELBERG.DE (LMail V1.2a/1.8a) with BSMTP id 1663; Mon, 13 Dec 1993 22:44:26 +0000 Reply-To: Mailing list for the LaTeX3 project In-Reply-To: Your message of Mon, 13 Dec 93 15:59:14 LCL Date: Mon, 13 Dec 93 12:17:36 MST From: Grant Gustafson Sender: Mailing list for the LaTeX3 project To: Multiple Recipients of Subject: Re: ideas for floats (forwarded) Status: R X-Status: X-Keywords: X-UID: 1203 > %> 1. A figure option to place a figure on the current page directly > %> after its citation. This implies that a figure citation exist which > %> marks a page and a page location where the figure is to appear. Dear Mike, Currently, LaTeX puts a figure where the float algorithm says to put it. The decision depends upon many parameters and it is complex. The settings used by the popular styles [article, book, report] for these parameters cause the figures to be placed in other locations infra, with no easy control over which page gets the figure. After reading and understanding the float algorithm in latex.tex (5-10 hours) it is possible to use the float parameters to design a style that is more predictable in the placement of figures. But it won't be perfect. The basic knapsack optimization algorithm in latex.tex would have to have many more frills to insure predictability. Many mathematicians who write books and longer articles have come to terms with the float rules and devised their own solutions. The faculty and graduate students here in my department (UofU Math Dept) use TeX exclusively and I can document numerous clashes with LaTeX figure and table environments. In the end we solve the problems, but the solutions are always hacks that can be broken later on by a spelling correction or a change in citation. > %> > %> 2. Item 1, but the figure goes in a particular place on the page: top, > %> bottom, margin or here. > %> > %> 3. Failure algorithm. If the figure cannot be placed where it is > %> desired, say page 10, then the page markup system creates page 10a, > %> 10b, 10c, etc, and dumps the troublesome figures onto these pages. > %> This is easy to implement in LaTeX and an acceptable kind of > %> failure resolution. Everyone consulted dislikes the mystery page of > %> floats that appears later on God Knows Where (application of the > %> float algorithm in LaTeX). Evidence: Lamport advises people NOT to > %> report this "error". > %> > With respect, isn't this *exactly* what LaTeX does now? It does for me, > anyway! There are lots of examples where LaTeX's algorithm works. That is why Lamport released it: it works often enough to be useful! There are lots of examples where it doesn't work as expected [but it does work as documented in the LaTeX manual and in latex.tex!]. In 7 years of sporadic updates of the univ of utah thesis LaTeX style sources many significant examples have appeared. A recent example: Tang's PhD thesis, 300 pages of LaTeX with over 100 figures and tables. Nelson Beebe spent over 100 hours helping Tang with the figures. A significant trouble for Tang was figure placement and the float algorithm (maybe it still is, he is supposed to submit a final copy in December). Communication of the actual problem with floats seems to be a difficulty. There are some simple test files (<200 lines) that can illustrate the problems better than I can do in words. If you are really interested, Mike, then I'll email an example to you personally. Best regards, Grant *================================* | Grant B. Gustafson | | 113 JWB, Dept Math, U. of Utah | | Salt Lake City, UT 84112 | | (801) 581-6879 | | email: gustafso@math.utah.edu | *================================*