Re: XSL Requirements (was: Microsoft extensions to XSL)

Chris <chris@xxxxxx> wrote:


>Oren Ben-Kiki wrote:
>
>> As Didier PH Martin (mailto:martind@xxxxxxxxxxxxx) correctly pointed out,
a
>> pattern matching language is never sufficient by itself to do general
>> structural transformations.
>
>Right. But then, XSL is not intended as a general structural
>transformation language. Its a style language that includes the ability
>to do some transformation as part of the styling step.


The situation as I see it is as follows (please correct me if this is
wrong):

- CSS and XSL (as originally intended) are meant to allow separating content
from formatting in  documents viewed in the browser.

- CSS (as it stands today) is based on the assumption that we begin with an
HTML document, which needs to be "annotated" with formatting information.
Structural transformations are minor if exiting at all.

- XSL (as it stands today) is based on the assumption that we begin with an
XML document, which needs to be transformed into HTML _and_ have formatting
information attached to it. Structural transformations are therefore major.
There is an assumption that the XML structure is not "too far" from the
resulting HTML structure, so the transformational capabilities are not
"complete".

- In practice, (current) browsers support CSS and not XSL. (Some) XSL
support is expected in the next generation of browsers, though.

- There is a real need for a standard mechanism for transforming XML into
HTML (and other languages). The W3 organization (currently) has no proposed
solution to address this need.

- XSL, being "the nearest thing" to a solution, is used to fill the gap.
This has caused people to push for XSL being a more "complete"
transformational language, regardless of the original intent. There are
several competing solutions, some based on CSS.

So much for the facts. I hope I got them right. Now for some opinions:

- CSS is more popular then XSL since it is seen as an evolutionary step
which fits in the HTML framework, while XSL is seen as a revolutionary step
(switching to XML as an input language). While this public image may not be
100% correct, it has enough influence that I wouldn't bet on XSL making it
as a stylesheet language.

- The XSL draft is being pushed in two directions and faces the danger of
becoming an unsatisfactory stylesheet language _and_ an unsatisfactory
transformation language. The first part would result from implementations
which view XSL as a transformational language (as most do today), and the
second part would result from designing XSL as a stylesheel language (as is
the current intent).

- While there are competing transformational solutions, all (that I am aware
of) are "corrupted" by being overly concerned with formatting issues. The
"right" solution is to have defined formatting languages (for example: HTML
+ CSS, a mathematical formulas formatting language, a 2D/3D graphics
formatting language, etc.). In addition we need to define transformation
language(s), which convert an input language (XML or HTML) to an arbitrary
(combination) of formatting language(s).

- Just like we have more then one formatting language, we may end up with
competing transformation languages. This is no excuse to tie particular
transformation languages with particular formatting languages (as is being
done today).

- If we'd insist that languages should be for "formatting" XOR
"transformation", then CSS should be moved to the "formatting" part, and
developement of transformation abilities within it kept to a minimum. XSL
should be broken into two parts (XTL and XFL). I don't see much future for
XFL if it is separated from XTL - which is a good indication of why this
procedure is necessary. Spice might be disentangled from CSS and become a
viable transformation language. There's also STTS...

And the most important point:

- The W3 organization should address this issue.

What needs to be done to get the attention of the W3 organization? I'm
thinking of conducting an informal survey (what would be the proper forum)
on this issue; if the results are favorable, that would probably get their
attention. Is that acceptable to the list maintainers?

Share & Enjoy,

    Oren Ben-Kiki


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