RE: Netscape Support for XSL - client vs server rant (Leventhal's articles)

Subject: RE: Netscape Support for XSL - client vs server rant (Leventhal's articles)
From: "Paulo Gaspar" <Paulo.Gaspar@xxxxxxxxxxxx>
Date: Wed, 24 May 2000 12:31:38 +0200
From: owner-xsl-list@xxxxxxxxxxxxxxxx
[mailto:owner-xsl-list@xxxxxxxxxxxxxxxx]On Behalf Of Jeff Greif> In the
various discussions of levels of support in IE5 and Mozilla, I didn't

> notice any mention of last year's opposition to supporting XSL within the
> Mozilla development team.  See Michael Leventhal's articles "XSL
Considered
> Harmful", in
> http://www.xml.com/pub/1999/05/xsl/xslconsidered_1.html , which asserted
> that XSLT was an ugly, unusable language, confusing and delaying the
> adoption of CSS, and that anything it could be used for could be better
done
> with procedural access to the DOM and use of CSS.I bumped on that article
some weeks ago.

It got me a bit stressed... so, I think I am going to use this opportunity
to let it out of my chest:
=;o)


For web browser use I tend to agree with Leventhal's when he says that CSS
can do what the XSL FO's do:
  - CSS sure seems to be doing the work on the more "graphic" side;
  - It is not clear to me that (for Web use) FO's would save enough on
    development effort in order to compensate learning them, much less to
    implement them.
However, this is what "I TEND" to believe. Since my reading of the XSL
standard about the FO's was superficial, I am not sure about this.


However, on the XSLT side, I think that Leventhal's really didn't know what
he was talking about. Even after superficial reading, the advantages of XSLT
over scripting on XML manipulation were already obvious to me. That's how I
immediately decided to go on with investing (a lot of) time on learning it.

XSLT is not a magic bullet ideal for all the XML manipulations. Sometimes it
is much better to use scripting/programming. However, XSLT style sheets are
specially suited to:
  1. Template-like work defining an (X)HTML layout for (XML) data;
  2. Other simple transformations better defined in a "declarative" way.
(1 is the most usual specific case of 2.)

XSLT clear advantages (from my experience) are:
 - It is much easier to imagine/write a XSLT template than an equivalent
   script;
 - You will find much more Web Designer's able to understand and customize
   the XSLT template then for customizing an equivalent script;
 - XSLT browser side functionality will by much easier to accept than
   scripting functionality in those companies suffering from Browser
   Security Paranoia.


However and again, XSLT is not a magic bullet for all cases. Scripting can
be the best solution in many cases. DOM manipulation with some XPath use can
work wonders.

(The idea of having scripting inside XSLT templates sounds specially
interesting to me)


What disturbs me most on Leventhal's attitude is the arrogance.

One should investigate things a bit further before making an absolute
statements.
(And after that I would still hesitate to make it too much of an absolute
statement.)


In the end, NO ONE knows it all.

Programming paradigms are changing all the time. You just needed to follow
the evolution of Object Oriented techniques during the last years to
understand that - some years ago with was "program to inherit" with huge
monolithic class hierarchies like the NIHCL and now its the Gang of Four's
"Design Patterns" cooperative objects style.

Let's keep an open mind and learn, learn, learn...


Have fun,

Paulo


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


Current Thread