User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.8.1.12)
            Gecko/20080213 Thunderbird/2.0.0.12
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID:  <4947FAAE.3090000@arcor.de>
Date:         Tue, 16 Dec 2008 19:59:58 +0100
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From: Stephan Hennig <mailing_list@ARCOR.DE>
Subject: Apache FOP approaches to  float placement and page layout
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
Precedence: list
Status: R

Hi,

searching the web I've been pointed to a Google Summer Of Code 2006
project that tried to teach float placement to the Apache Formatting
Object Processor:

<URL:http://wiki.apache.org/xmlgraphics-fop/GoogleSummerOfCode2006/FloatsImplementationProgress>

Another page on the same wiki points to lots of documents discussing
FOP's page layout approach:

<URL:http://wiki.apache.org/xmlgraphics-fop/DeveloperPages>

I do not know how well the LaTeX3 team knows about other software and
its approaches to typesetting problems.  It seems to me FOP people adapt
many of TeX's approaches, but augment them by their own ideas.  (The
same is true for hyphenation in OpenOffice, BTW.)

To quote from the discussion of float placement:

> In his thesis (see below), Plass defines a badness function for
> measuring pagination quality that makes the problem NP-hard. In this
> article, another function is used which in the authors' opinion
> better reflects a human's judgment, and for which there is a
> polynomial solution. In short, they use the number of page turns
> necessary to read the document, that is, the total number of pages +
> the number of page turns caused by figures appearing on a different
> page from their citations. Plass used the sum of the squares of page
> differences between figures and their citations.

I don't know if LaTeX3 can learn anything from FOP's float placement
approach.  At least it seems to be an extensive discussion of the matter.

Best regards,
Stephan Hennig