[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: DOCBOOK-APPS: Re: Future of jadetex (Was: First Open SourceDocumentation Summit at the O'Reilly Open Source Convention




 > 1) currently, DocBook -> professional print is not convenient

"convenient" is a vague word. clearly, many many people do
professional work with DocBook, from O'Reilly downwards.

 > 2) the base of a possible new tool should be XML and no more SGML
I think so too

 > 3) the new tool must handle many languages specificities correctly
sure

 > 4) none of the current free' processing tools meet the above points or
 > even took incompatible directions.

I probably don't agree, but no matter

 > 4) LaTeX gives professional prints for *many* languages and uses free'
 > tools

well, not so. LaTeX has many problems (come to TUG 2000 next month and
find out!), does not deal with all language, and not all of it is free
(eg the "french" package). a lot of what one traditionally associates
with languages in LaTeX (Babel) is not appropriate in this environment
anyway

 > proposal:
 > 
 > 1) Define the mean to pass from DocBook to LaTeX:
 >   - perl/Python
 >   - DSSSL
 >   - XSL
to write the simple transformation from DB to LaTeX would be easy in
XSL, and would allow for multiple platforms and degrees of
embedness. DSSSL is a no-no (lack of multiple implementations), and
Perl/Python is too low-level. IMHO.

 > 2) Solve the stylesheets problem (depends on point 1)
 >   - use current .dsl ones

sorry, but you cannot have it both ways. if you use the .dsl styles,
you immediately dip well below the LaTeX radar, and have to create
low-level TeX like JadeTeX. Sure, thats an option, and not a bad one,
but is it what you want? you'd lose the high-level stuff you seem to
want LaTeX for

 >   - use LaTeX .sty ones
and that ties you to an undocumented style language...

 >   - develop a [dsl2xsl, dsl2sty] script?

nice idea. you could write a high-level Docbook to LaTeX converter
(map <foo> ..</foo> to \begin{foo} ..\end{foo}; you could do that in a
single XSL template, see appended; just needs a template for
attributes),  and then convert the .dsl to .sty?
but you'd just end up rewriting JadeTeX. In fact, its exactly what
xmltex does anyway!

dsl2xsl? what would that buy you?

In case you think I am just being negative, my suggestions are

 - switch to XSL FO and make PassiveTeX work (maybe with a
    pre-processor)
 - process the "fot" output of Jade with TeX
 - convert the "fot" output to TeX
 - write a new TeX backend for Jade with a lot more intelligence in it

I have long felt that the last of these is the only effective way
forward for DSSSL, and what we should have started a long time
ago. The problem is that it needs someone who understands the way Jade
works, is a good C++ programmer, and who has a good understanding of
TeX. At least one such person exists (Matthias Clasen, to name names),
but does any one want to volunteer? 

Sebastian


<?xml version='1.0'?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                version="1.0">

<xsl:output method="text"/>

<xsl:template match="*">
 <xsl:text>\begin{</xsl:text>
   <xsl:value-of select="name()"/><xsl:text>}</xsl:text>
    <xsl:apply-templates/>
 <xsl:text>\end{</xsl:text>
   <xsl:value-of select="name()"/><xsl:text>}</xsl:text>
</xsl:template>

</xsl:stylesheet>


--  
To UNSUBSCRIBE, email to ldp-discuss-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org