CSS for transformation

[Paul Prescod <papresco@xxxxxxxxxxxxxxxx>]

Jelks Cabaniss wrote:
> 
> > I don't know how I could have stated this any more clearly, but: if CSS's
> > model was sufficient, they would have just used CSS's model and there
> > would have been no two things to harmonize, right? I didn't claim that
> > CSS's formatting model was miles away from XSL's. I merely said that it
> > *was not sufficient* (in the eyes of the people who decided not to use
> > it).
> 
> Presumably because in the Holy Trinity of XML, XLL, ???, CSS didn't begin with
> an "X".  

I don't know that there is any real benefit in trading conspiracy
theories. After an initial scan, the XSL formatting model looks much more
complete than the CSS model. That doesn't mean that the CSS model is
wrong, or stupid or was a waste of time to create. It is simply not
sufficient. That's the nature of progress. Usually it isn't contentious to
say that older technologies need to be updated to solve new problems.

> Also, there's a bandwagon passing by and you have to be wearing
> pointy-brackets to get on.  I do wonder what Postscript is going to look like
> with pointy-brackets ...

There have been SGML-based page description languages for several years.
You can look at SPDL, for example.

> I don't know how the Style and Activity pages could state it more clearly:
> there's no "CSS's model", just the W3C Formatting Model -- for XSL *and* CSS,
> the latter of which will work with both XML *and* HTML documents (and with a
> much simpler syntax, IMHO).

There is a "CSS formatting model". The CSS model is defined in the CSS
specifications. A version of this model existed before XSL was even a
twinkle in anyone's eyes. Eventually there will be a W3C Formatting model,
but that is the future, not the present. For now, I stand by my (I
thought, uncontroversial statement) that the CSS model was not sufficient
and would need to be extended (and perhaps modified) in the creation of
the XSL formatting language.

> Though CSS currently does not yet really do transformations of elements,
> attributes, and content, see http://www.w3.org/TR/NOTE-stts2.html.en for a taste
> of what's apparently on the way...

Interestingly, this language has a little less expressive power than
architectural forms from the SGML world. In my experience, this is not
enough to accomplish formatting tasks. Fundamentally, you need to be able
to wrap and reorganize elements. I believe that both mechanisms are
weakened by their need to fit into a constraining pre-defined syntax. A
CSS like language that was as expressive as the XSL transformation
language would look about as complicated as the XSL transformation
language, and not much like current day CSS.

 Paul Prescod  - http://itrc.uwaterloo.ca/~papresco

Bart: Dad, do I really have to brush my teeth?
Homer: No, but at least wash your mouth out with soda.


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list