DOCBOOK-APPS: html-xslt comments, questions, requests

[Vincent-Olivier Arsenault <vincent at up4c dot com>]

Hey all!

First I'd like to congratulate all the contributors to the XSLT package, 
because it is a very, very handy tool!

I'm using the HTML templates of the 1.52 version of the XSLT distro.

Here are my user notes..


########################
# 1.  <div> generation #
########################
Looking at the resulting code, I find that there are tons of <div>s 
everywhere (some of them have no attributes).

I am wondering if it could be possible to have an elegant way of turning 
it off without rewriting about 40 big templates (maybe with a 
"make.block.div" parameter containing the docbook element(s) to be 
transformed into <div>s)...

Anyways, it's because <div>s in html also carry presentation 
instructions that are irreversible using CSS: for example, they won't 
flow around floated elements no matter what you do (unless I missed 
something)...


#########################################################
# 2.  chunk names and URLs for permanent web addressing #
#########################################################
I think having the chunks named in a hierarchical way (using 
directories) would be a great feature for people who would like to 
render a complete web site out of a single docbook document, because :

1. URL should represent the structure of the information so that it is 
easy for the user to know where he is.

2. URL should be descriptive of their content.

Also, link paths and chunking path should be generated differently. For 
example, in the case of an href,

(a) "/myBook/myPart/myChapter/"

is better than

(b) "/myBook/myPart/myChapter/index.html"

Because while they will both get you the resource you want, (a) allows 
the webmaster to change to another server side technology (that would 
change the file name to "/myBook/myPart/mychapter/index.php", in the 
case of PHP) without having to redirect or to issue 404. Another reason 
for that, more obvious, is that (a) is smaller and that the file part in 
(b) is not descriptive of its content.


###############
# 3.  preface #
###############
I would like the preface to be part of the book's chunk. Would that be 
semantically correctly, is there another way to get text on a book chunk?


################################
# 4.  ToC in a separate chunk  #
################################
MMM, would it be great to have the possibility to put the table of 
content in a separate chunk and to be able to specify its path? Or maybe 
it can be done already?

;-)


And if there is code to contribute to do that, I'm willing to help. I 
just thought it would be nice to have some directions from the authors 
and also indications on whether those features are OK to include in the 
distro.


have fun,

vincent-olivier

ps : Has anybody tried Adobe Framemaker 7? It supposedly supports XML 
docbook. Our copywriters would need a "m$-world-like" app in order to 
have a full docbook publishing pipeline. Any related experiences, advices?