On Tue, 7 Dec 2004 will@phototropia.org wrote:

  [...]

> 1) There is so much talk about IE not properly rendering css, that it
> clouds the fact that css is now the best keep secret that everybody knows.
> The w3c web-site claims that css is not the thing of the future and that
> xslt is the thing to use.  They explain their reasoning based on xslt's
> capabilities/features.  They claim that css works well for styles for one
> page, but lacks the features that xslt can offer to format an entire
> document.  Many even in the technical realms tangential to web-application
> design believe that css is cool, but never really happened because of the
> browser issue.  This is what led me to lean towards xslt as I like xml.

Personally, I think of XSLT and CSS as two non-related technologies. One
is for converting one form of XML to another (or something XML-like),
while the other is for instructing the viewing-application (ie. browser)
how to render a particular variation of XML. In the XSLT world, you have
XSL:FO (which isn't exactly XSLT, but "the other half" iirc). See my
next point about that...

>       However, quantums of professional web-app designers don't like
> working with xslt.  I don't want to over-extend myself quoting
> other mail-archives/comments, but the jest of their dislike evolves
> around the yield of features gained vs. the syntaxical
> complexity/(design?)

As I understand it, in order to use XSL:FO to do the same job you can
easily do with CSS, you need a big lump of slightly messy XML, littered
with complicated XPath and XLink expressions. Compared to CSS, it's
syntaxical hell, not to mention verbosity. The features gained by XSL:FO
compared to CSS are negligble, the real difference is between XSLT and
CSS, which is natural, considering they are tools for very different
jobs.

  [...]

>> Most people dream that they can solve with xml and xslt a big problem:
>> One source for print an online information. But then you can't use HTML
>> tags,
>> then you need to use docbook. And again that's too complicated.
>>
>> I have a simple script which translates some HTML tags to latex. That's
>> enough
>> for me needs.

Didn't react when I read this post the first time, so I'll hijack onto
this one. :)

In my work on my master thesis, I use docutils and reStructuredText. RST
is simple and non-intrusive, yet yields (atm) output to both HTML and
LaTeX. There are output-writers in the works or in beta-stages for
docbook, PDF and a few other formats. It's worth taking a look at before
you start creating your own XML+XSLT nightmare to solve an already
solved problem. :)

--
Morten
Your friendly neighborhood Atheist.