Re: Venting

Subject: Re: Venting
From: Paul Prescod <paul@xxxxxxxxxxx>
Date: Wed, 03 Feb 1999 23:29:25 -0600
James Clark wrote:
> This seems to be rather overstating things to me.  The fact that the two
> languages are defined in one spec doesn't affect the languages
> themselves one iota. 

I don't believe that.

Let's start with the question of whether XSL is one language or two. The
spec. says:

"XSL is a language for expressing stylesheets."

One language. Not the two that your sentence implies. There's our first
problem. Our intuitive understanding is out of step with the spec.

I don't think that the two languages even have names. I do not believe
that the word "transformation language" is in the XSL specification.
Neither is "formatting object language". "Formatting object vocabulary" is
in the specification but a vocabulary is not a language. A language would
have a DTD of some sort (which I would expect the formatting object
language to have some day).

What exactly do instances of the unified "XSL" language do? Arbitrary
transformations? It seems not:

"An XSL stylesheet specifies the presentation of a class of XML documents
by describing how an instance of the class is transformed into an XML
document that uses the formatting vocabulary."

What about processor conformance?

"When the result tree uses the formatting vocabulary, a conforming XSL
implementation must be able to interpret the result tree according to the
semantics of the formatting vocabulary as defined in this document;"

This makes very fuzzy the idea of conformance. Do XT or IE 5.0 "interpret
the result tree according ..."? Are those products legitimate in calling
themselves XSL implementations?

>  The specifications of the two languages are
> cleanly separated: the transformation language is in section 2 and the
> formatting DTD is in section 3. Whether or not you make separate
> physical documents out of the two sections doesn't seem a big deal to
> me.

It is a big deal because people come to a language called "XSL" with
expectations. When I show them XT with DocBook output they say: "I don't
see anything that remotely resembles style application." Well, I say, 

"Style application is ONE thing you can do with XSL but only one. It isn't
REALLY a style language at all. Or, to be precise, it has two parts and
the two parts together are a style language but if you throw away one part
you end up with something which is much more general and in many ways much
more powerful. Sorry. I'm confusing you. Let me back up. You take the XSL
style language and you throw away the style part and you get this cool
transformation language. What's it called? Well, it doesn't really have a
name. It's XSL without the style parts." 

I've spent my life explaining that 

 * LINK has nothing to do with hyperlinking, 
 * "DSSSL" and "Grove" are not an acronyms that it is useful to expand, 
 * "object oriented" is actually more "object based" than "object based",
 * JavaScript is not a subset of Java
 * "insignificant whitespace" is not, 
 * "XML text" is not the text in an XML document, 
 * an XML document can be "invalid" but proper, 
 * "mixed content" is not necessarily very mixed
 * Non-SGML entities can point to SGML
 * element declarations are not

Have some pity on an old educator. Do not introduce another one of these
monstrosities into our vocabulary. 

 Paul "XSL is not." Prescod  - ISOGEN Consulting Engineer speaking for
only himself

"Remember, Ginger Rogers did everything that Fred Astaire did,
but she did it backwards and in high heels."
                                               --Faith Whittlesey

 XSL-List info and archive:

Current Thread