Re: DOCBOOK: docbook vs latex

[Doug du Boulay <ddb@R3401.rlem.titech.ac.jp>]

On Tuesday 03 September 2002 20:12, Michael Smith wrote:
> It really does seem like math markup might be a unique case.
>
> For one thing, discussions related to mathematics markup/rendering turn
> up pretty regularly on the docbook and docbook-apps list. But I can't
> remember discussions about markup for many other non-computer-related
> fields ever taking place on the lists. I mean, I don't remember seeing
> suggestions for adding, say, "biology" or "chemistry" markup to DocBook.
> Another thing: math markup seem unique in that it has specific and
> unique processing expectations/ rendering requirements associated with
> it. I don't think most other kinds of discipline-specific markup have
> such specific processing expectations associated with them.
>
>   --Mike

On Tuesday 03 September 2002 23:34, Peter Ring wrote:
> Some non-mathematical applications with highly specific rendering
> expectations:
>  - music
>  - chemistry

I wouldnt be suprised if there wasnt already a dozen variants on MusicML.
Music being a relatively restricted language I imagine the 
rendering of notes to graphics and audio could be achieved relatively easily
(in the sense that MathML can be rendered to math easily)

On the topic of chemistry, ignoring the demonstrated  lack of demand
for basic things like <chemicalformula> etc, 
the highly specific rendering mentioned above would probably be 
for things like chemical reactions, phase diagrams,
and rendering of  formulae into structural diagrams. 
The latter generally requires a significant
amount of extraneous detail to compute. You could always 
include a preconstructed image, much like equations, but then why
bother with extended markup at all? To document chemical structure
properly you would probably need to turn the whole document into some kind of 
chemical data base entry. 
This been done already using the 
crystallographic Information File (CIF) (basically a poor mans XML) which is
a single data base file containing raw text as data alongside experimental 
and structural data. The whole file is parsed, data is checked and text is 
rendered into html, pdf and hardcopy editions of the journals
See http://www.iucr.org  for details.

The sad thing is that the online text is relatively lifeless. I think the 
bibliographic entries are currently cross referenced across journals
to some degree, but otherwise the text just sits there. i.e. there are no 
hyperlinks to glossaries of terms or popup links to related structural
families. Viewing of the structural model can be done via external browser
plugins or some such rather than being an integral part of the document using
javascript and java.

If DocBook and associated tools provided the hooks to get all that kind 
of thing happening, it might actually stand half a chance of getting 
accepted into mainstream chemistry publishing.

There is also a competing technology to CIF called the 
chemical markup language     http://www.xml-cml.org/
for marking up chemical data, but not documentation.
Check out the CML DOM link on that page for a list of 
XML elements contained therein. Maybe they could be linked 
to DocBook as modules  similar to MathML.
CML is largely the work of one or two people so it could evolve
rather rapidly and is perhaps not officially sactioned. 
On the other hand the developers did have high aspirations that may 
not have come to fruition yet: see http://www.ch.ic.ac.uk/rzepa/cc96/


On Tuesday 03 September 2002 23:34, Peter Ring wrote:
> Somewhat mathematical applications with slightly less specific rendering
> expections:
>  - data and system modelling, UML
>  - semantic networks, conceptual graphs
>  - graph theory

So now you have MathML, MusicML, ChemML and  UML
(+ GraphML? + LiterateProgML? +SVG +...) Can a modularized DocBook
provide hooks for all these plus more? 

I guess Matt's last email and this message from the archives 
pretty well hits the nail on the head:
http://lists.oasis-open.org/archives/docbook/200202/msg00087.html

Doug