Bug List of OpenJade As I Know

[intron at intron.ac]

1.HTML
  1) If there is a " " in index term like:

           <indexterm><primary>A B</primary></indexterm>

     the output HTML will include "A&nbsp;B". This bug affects
     at least Chinese.

  2) If index is inside "<part>...</part>", the link "Prev" at the
     head of doc-index.html will point to doc-index.html itself
     (This bug may concerts DSSSL stylesheet). See:

     http://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/doc-index.html

2.TeX
     Chaotic bookmarks in PDF (Error ancestor-descendant relationship.
     This bug may also concerts DSSSL stylesheet):

                    Chapter 1 Introduction
                        1.1 Developing on FreeBSD
                            1.2 The BSD Vision
                            1.3 Architectural Guidelines
                            1.4 The Layout of /usr/src

     NOTE: PDFs on ftp.freebsd.org are built with Jade, not OpenJade.
           In this case, no bookmark is properly generated at all.

3.RTF
  1) OpenJade mis-marks some languages other than Japanese with
     Japanese code page ("\fcharset128").
  2) OpenJade takes it for granted to use the font "WingDings"
     of Microsoft Windows (hard coded) for the numeric callouts
     circled 1, circled 2, ..., which results in OpenOffice's
     failure to display callouts without the font "WingDings".
  3) OpenJade cannot embed images into RTF, though it is easy.

4.For all output formats, OpenJade lacks enough support for all
  character sets covered by Unicode. This bug has affected Greek
  FreeBSD documents, though ISO 8859-7 is a simple character set.

     Above all, Jade/OpenJade is no longer an active/vivid project.
Even Jade's original creator, James Clark, has moved his interests
onto other issues. We can see "James Clark" on XSLT/XPath
specifications:

     http://www.w3.org/TR/xslt
     http://www.w3.org/TR/xpath


Hiroki Sato wrote:

> intron at intron.ac wrote
>   in <200604021732.k32HWiqn099718 at zhao.intron.ac>:
> 
> in>      Actually, if FDP is transfered into a XSL system, only a few
> in> FreeBSD-specific DSSSL style sheets need to be converted into XSL style
> in> sheets, and those SGML DTDs need to be converted into XML DTDs with
> in> slight changes. After all, most of style sheets for FDP are provided
> in> by Norman Walsh, either in DSSSL or in XSL.
> 
>  FreeBSD specific DSSSL stylesheet is not small, and the resulting
>  documents are different between DSSSL and XSLT.
>  I have implemented SGML->XML conversion in NetBSD Documentation
>  Project (you can see htdocs/share in their repository), but
>  from my experiences, that work is more problematic than one you think.
> 
> in>      In fact, Jade/OpenJade's bugs (mainly around its tex-backend) are
> in> only minor. TeX's bugs and the glued relationship between Jade/OpenJade
> in> and TeX is the substantial of problems. Sometimes, we cannot find
> in> a way out for Jade/OpenJade to let TeX consider it as "correct", and
> in> maybe even both two sides of a problem are taken as "wrong" by TeX.
> 
>  Please show me some concrete examples?  I am still not sure
>  what your problem is.  "Which is correct" is always determined
>  by the Jade + JadeTeX processing model and such problems should be
>  discussed with the authors.  I am interested in it if there is
>  a critical problem that we cannot solve now in the current framework.
> 
> in>      At least today XML/XSL is the developing trend rather than
> in> SGML/DSSSL. So many efforts are focused on XML/XSL in open source
> in> community all over the world. XML/XSL system supports Unicode natively,
> in> which means all languages including your Nipponese may be well solved.
> 
>  Agreed, but the switching over is not straightforward.  I guess you
>  think it is relatively easy.
> 
> in>      Now MySQL's docments are typeset by Apache FOP, which have better
> in> quality than current FreeBSD documents (http://dev.mysql.com/doc/),
> in> even only in English. Their PDFs have commercial-level typesetting
> in> styles in some details: word-splitting, margin kerning and font
> in> expansion, just as LaTeX + microtype package do. It seems that XSL
> in> system is absolutely not a simple "it works" one, at least witnessed
> in> by 1727-paged "MySQL 5.1 Reference Manual".
> 
>  So, do you have a concrete plan to adopt such typesetting
>  framework?
> 
> --
> | Hiroki SATO