Re: CSS for transformation

Philippe Le Hégaret wrote:
> 
> Paul Prescod wrote:
> > I didn't claim it was a big problem. I didn't claim that the CSS model
> > could never have new functionality added. The truth is that the XSL model
> > will be largely based on the CSS model. It will essentially be the
> > formatting part of CSS3.
> >
> > > And if people wants an X, it's very easy to create the XCSS.
> >
> > Sure, but why?
>   Because I don't want to implement two differents style models
> in my application, one for CSS and one for XSL. I don't like
> the word "largely" or the word "essentially".

Well, if we take out the word "largely", we have:

"The truth is that the XSL model will be based on the CSS model."

Which I believe is still an accurate description of the plan. We can do
the same with "essentially":

"It will essentially be the formatting part of CSS3."

Which I also believe is the current plan. Does that make you feel better?

> > I've used more than a dozen transformation languages, and XSL is the most
> > well thought-out of the bunch. It is an *excellent* transformation
> > language, and it works beautifully.
>   No, for me it's not. I don't want to support an undocumented
> post processor in my XSL. I don't want to support a specific
> XSL engine just because this is the only one post processor
> for my output format.

So you would rather make XSL inappropriate for its main task of
transforming documents for display on the Web? Just so that you can use it
as a "report writer" or "converter"? There are lots of great report
writers and converters out there already. Use Python for your conversions
and leave XSL for hat it was intended to do.

>   Not agree with this. It's an another way to say : "You can't do the
> conversion with XSL, so write a program for this.". XSL can't define
> a namespace for all existing format.

No non-programmable language can convert (e.g.) an average XSL document
into (e.g.) Postscript. If you want a programming language, use a
programming language: there are several good ones out there.

>   You can program a lot of things with only an if, a for and a when. 

Yes, XSL can be used in complex ways. So can "make", or "SQL". But they
are not programming languages. Nevertheless, you cannot (e.g.) compute the
value of PI with XSL. Any real programming language would allow this. When
it is possible to write a Turing machine in Python, then I will agree that
it is a programming language.

 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