Re: syntax feedback

[Chris Lilley <chris@xxxxxx>]

Paul Prescod wrote:
> 
> "Simon St.Laurent" wrote:
> >
> > The W3C has made some very strange moves, but putting CSS and XSL into
> > direct competition is one of the strangest.
> 
> They are not in competition. I admit that XSL is much more complex and
> thus will not be as popular among low-end stylists as CSS will be. CSS's
> inventors admit that CSS is much less powerful and thus *cannot solve* the
> problems that XSL is designed to solve. 

They are certainly aimed at different markets; and they can also be used
in tandem - server side  transformation from data-like xml to
document-like xml followed by client-dise styling using CSS.

> XSL is the third standard stylesheet technology to be used by the generic
> markup world. The first one failed because it did not allow
> transformations. CSS was never even a contender because it did not do
> transformations.

So if the first was - what? Panorama stylesheets, FOSI ? and the second
is CSS, and XSL is the third, where does DSSSL fit in? According to your
logic, DSSSL was a runaway success because it provided not one but two
ways to do transformation.


> Our own tests demonstrate that they are NOT in contention. CSS people do
> not like XSL because they don't understand why it has added complexity
> because their problems do not require the added complexity. 

I'm a CSS person. I don't have a problem with XSL (I do have a problem
with inaccurate comparisons) and I do see the need for both the
XML-to-XML transformation part and also the need for a ich formatting
model with huge numbers of knobs to tweak, most of which are left at
their default settings most of the time.

It seems, however, that fols implementing or experimenting with XSL are
merely content to spit out extremely poor, presentational, inaccessible
HTML which, to my mind, takes us furmly back to the sort of grungy mess
that the Web was in a couple of years ago and which CSS was internded to
solve.

> If you do NOT
> need transformation, then there is no benefit in XSL's higher cognitive
> load, so CSS makes much more sense. So they stick out their tongue at it
> (as you have described).

Transformation and styling need not be tightly coupled. They can be, but
that is just one possible design decision.

> XSL people cannot use CSS except as an adjunct to XSL. If you need
> transformation, then you don't really have any choice about whether to use
> CSS or XSL and there is also no competition. CSS can't even handle a
> simple cross-reference (nor should it be extended to! It's good at what it
> does)

There are other transformation technologies that can be used, both
procedural and declarative. Personally I would far rather use a site
that provided document-style XML and CSS, rather than some site which
provided crappy table-based HTML "formatting"; knowing that somewhere on
the server (but not available to look at)  was XML source and and XSL
stylesheet would not give me warm fuzzies. It might as well be an Oracle
database, or anything.

> According to our own informal survey, these technologies are not in
> contention, direct or otherwise. They are complementary.

I agree that they are complementary. I would like to see some XSL
implementations that actually implemented a style language, though. I
know, there is FOP. I would like to see more.

--
Chris


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