RE: XSL considered harmful

Jim Whitehead (
Fri, 21 May 1999 12:29:03 -0700

> A remarkably fresh set of articles titled "XSL considered harmful"
> have been published by on [1]. Tim Bray distances himself only
> slightly, and Michael Leventhal is given prime bandwidth for his
> "Declaration of war" [2]. His argues that:

Well, of course he has been given prime bandwidth -- nothin' like a
controversy to drive up readership. I'm sure would give anyone
prime real estate for any kind of a controversial stance concerning XML.

But, what are the implications of this article for the openness of the W3C
process if parties feel a major debate like this has to be taken public, and
can't be resolved within W3C fora? Perhaps the W3C needs a forum for
evaluating the technical direction of the W3C, in addition to all the fora
for discussing issues once the technical direction has already been set.

> - XSL, a "sometime in the future" technology, full of beautiful (if
> vague) prognostications about its "power" and "richness", offers
> no useful improvement in capability over current and implemented
> full W3C Recommendations for stylesheets and transformation.

I just love this argumentation style (not). Start your sentence with a
bordering-on-ludicrous restatement of a strawman of an XSL proponent's
viewpoint, then end with a reasonable criticism, thereby lending credibiliy
to the strawman. Still, have to give credit where it's due -- this tactic
quickly adds lots of emotion to the issue.

> - XSL is the most hideous and unwieldy language imaginable and
> stands absolutely no chance of acceptance by the web community.


a) There have been plenty of hideous languages in the history of CS -- XSL
would have to be really, really bad to take the top prize. Is it worse than
Ada 95 and Cobol?

b) Language elegance doesn't necessarily have anything to do with standards
adoption. For example, C++ was adopted despite its many flaws.

> - XSL advocacy has blurred the focus of the W3C, by introducing
> competing standards for styling and transformation, and set back
> by at least two years the goal of vendor-independent, semantically
> rich "open information highway", by undermining support for existing
> standards such as CSS and the DOM.

Ah, I get it:

"open systems" + "information highway" = "open information highway" :-)

Plus, I suspect it's a simplification of matters to say that XSL being up in
the air kept MS from fully implementing the CSS standard. This is pure
speculation, but I suspect part of the reason MS and NS have had a hard time
implementing CSS is that their internal representations of HTML weren't
designed with this in mind, and hence implementing CSS would require deep
modifications to HTML processing, which is at the core of the browser.
Anyone I've ever known who has worked on abstract syntax trees has found the
work to be tedious, hard, and detail-oriented, and I expect adding CSS
support was no different.

However, the arguments in the body of the article do have some merit (and
are generally much more calm). It's just a shame Michael Leventhal had to
water down his overview points with all the rhetoric. To his target
audience of opinion makers, "XSL Considered Harmful" *is* a declaration of
war -- repeating this overtly in the subtitle just makes him seem less
confident ("hellllooooo, pay attention to me, yo!, I'm declaring war here
guys, HEY!")

Hakon, I'm surprised you didn't also send along your op-ed piece on the
subject, "Formatting Objects considered harmful", quoted in the article,
located at:

- Jim