Re: [xsl] disable-output-escaping

Subject: Re: [xsl] disable-output-escaping
From: Joerg Pietschmann <joerg.pietschmann@xxxxxx>
Date: Fri, 07 Dec 2001 14:45:20 +0100
Wendell Piez <wapiez@xxxxxxxxxxxxxxxx> wrote:
> At 07:43 PM 12/3/01, Mike Brown wrote:
> >My experience with developers who are well-versed in other languages but
> >not in XML is that they don't see any compelling reason to maintain such a
> >markup/text distinction. There is really nothing wrong with wrapping HTML
> >in XML; this is not indicative of poor application design.
> I agree with this: there's nothing wrong with it in the general case, and
> whether it's poor application design really depends on other factors.

Hmm. XML has its roots in a language for writing documents for
publishing purposes. For quite a few people here, the focus is
the document, a bunch of text often written by a human with some
more or less helpful text editor, with some markup sprinkled in
in order to get a certain appearance in the published form.
Some other people have shifted their point of view, placing the
focus on the structured information contained in an XML document
(with some special weight on "structure"). From this point of
view, a XML document, a stream of SAX events or a DOM tree are
simply different ways of expressing the same information. In this
view, there is no difference whether you pass the information as XML
document, as DOM tree or SAX event stream between processing stages. 
You must decide in this context whether "wrapping HTML in XML" means
"embedding HTML as structured information" (as subtree), or whether
it is a plain string. You can't expect the latter to be interpreted
by processors which expect to be fed with any representation of the
information (like a DOM tree). You need a parser to process a text
with markup into a representation which such processors understand.
Distinguishing processors which transform the information from
one representation into another (for example parsers and serializers)
and processors which actually do something with the information (like
XSLT processors or HTML renderers) helps to deal with complexity.
When this distiction is blurred by relying on serialization of the
information between processing stages, it is more often a symptom of bad
application design than not.

> >  The problem is
> >merely that they view their intended output as a pastiche of strings, and
> >XSLT as instructions for generating those strings, whereas they should be
> >viewing their output as the serialized form of what is essentially a
> >single XML document that was created one abstract node at a time.
> This is a nice way of phrasing exactly what I mean. Markup is not text:
> instead, it is a notation for or representation of something extra-textual
> - -- the structure of abstract nodes.

I agree to both points, with the hope of having it made even more clear.
> Anyway, this is edging into not being an XSL topic anymore

Oops, sorry. Couldn't restrain myself...


 XSL-List info and archive:

Current Thread