[
Lists Home |
Date Index |
Thread Index
]
I agree that for the subset of XML files that are XHTML, better
CSS is better XHTML. I concede the likelihood that more
HTML is being produced than any other markup language.
I would counter that separation of CSS formatting by annotation
and XSLT transformation is a good thing. I don't see that lack
of improvements in CSS can be blamed on the success of transformation.
What I concede is that there are documents that look not so good
because transformation to HTML is the only tool applied by some
set of developers, and that improvements in CSS would improve
the opportunity further. Confusion in which tools to apply
is a training issue. Not improving tools themselves is a
different issue. I don't know if it makes sense to claim
that XSLT is at fault for the failure to improve CSS, or
if the CSS users aren't able to influence vendors for other
reasons (belief that it is good enough might be one). It
seems to me that there is a competitive opportunity here.
I don't consider XSLT a stylesheet. XSL-FO is a stylesheet.
The semantics of style are conflated with transformation and
that is a heritage issue from DSSSL. Spilt milk. I didn't
like it then either. Hyperlinks were dragged into it too.
It would be fun to wind back the clock and see if James et al
would have done the work they did had they been told that
transformation but not formatting objects were wanted. However,
they separated these into XSLT and XSL-FO and that should
be sufficient. Confusion of users may be epidemic and in
some individual cases, political, but show me an application
language without politics and I'll show you a dead language.
I am told that X3D and SVG are different languages for different
applications, but I'll bet my next paycheck that if X3D becomes
more noticeable, the urge to make SVG-3D will go mainstream. It
isn't necessary, but it is very human.
My costs aren't going up per se but unless I don't bother to
acquaint developers with both CSS and XSLT, my rendering
quality is not optimum.
XSLT is absolutely necessary. CSS is excellent to have.
Better CSS would be more excellent.
len
From: Frank [mailto:frank@bunter.therichards.org]
On Wed, 2003-07-09 at 10:16, Bullard, Claude L (Len) wrote:
> That would suggest that the problem today is not
> technical or political, but a matter of education.
> There simply are XML languages for which CSS is
> not useful or available. There are applications
> of CSS which are simpler and more elegant. Without
> discrimination, one will stumble.
But there is also one big application -- documents on the web, for which
good CSS support would be a major improvement. While there is no
_technical_ problem with using two stylesheets, it is a cost.
Like Simon, I largely deal with documents that are basically human
readable as created. Except for the TOC and indexes (and indexes tend to
be replaced by links and searches anyway), they merely need to be
prettyprinted. Without support for generated text and counters my
clients have to have two stylesheets rather than one. I would bet that
half the XSLT I deal with would go away with those two enhancements.
Further, half the remainder would also go away: There's lots of other
code in the applications that could transform (or create right the first
time) the XML for CSS styling, if we didn't have to have the XSLT
anyway.
I accept that for most xml developers it's not a big problem. They seem
to mostly be moving information between computers. For those
applications which have a human at the beginning and the end, it's a
pain in the neck and a pain in the wallet.
|
