W3C home > Mailing lists > Public > www-style@w3.org > August 2002

Re: Serving generic XML

From: Coises <Randy@Coises.com>
Date: Mon, 19 Aug 2002 02:58:57 -0700
To: www-style@w3.org, www-tag@w3.org
Message-ID: <5jb1mu0gtjhbu4hpc8hqj9d4nmf72he2g2@4ax.com>

[Mon, 19 Aug 2002 02:08:59 EDT] Svgdeveloper@aol.com:

>Presumably, then, you are wholly antagonistic to one of the foundational 
>goals of XML 1.0 (at least as I understand it):
>"Its goal is to enable generic SGML to be served, received, and processed on 
>the Web in the way that is now possible with HTML."
>I would be interested to know how others read the quoted sentence. Is it W3C 
>double-speak which essentially means nothing but sounds rosy? Is it a 
>commitment/ambition which W3C has hitherto failed to realise? Or what?

The problem comes when you imagine that XML can be expected to provide
any new value to the web, to authors or to users.

The advantages of XML accrue to software developers.  As I understand it,
one might think of it as "SGML Light."  It provides a consistent, powerful
yet reasonably simple data model for structured documents.  This means
software vendors can build generalized processing tools (particularly
renderers) which can be reused in a variety of specific contexts.

I'm not denying that this may be a useful thing.  I just wince when folks
seem to credit XML with quasi-magical properties that it does not possess.

For any particular XML "dialect" to be used effectively, its semantics must
be available to the developers of applications that will process it.  This
is fine for files with a relatively limited "audience" (like *.asx files
for Windows Media Player, which appear to be a species of XML); but for
documents that anticipate general distribution (like web pages), the
problem of defining the semantics of a particular XML application is
virtually identical to the problem of defining a markup language standard.

Consider the XHTML 1.0 standard (<http://www.w3.org/TR/xhtml1/>): it
doesn't define semantics; it just borrows them from HTML 4.  After all the
hoopla about "the semantic web" --- XHTML 1.0 is *semantically*equivalent*
to HTML 4.01!  Make the tags lower case, add a forward slash here and
there... it's the same thing.  What's changed is that the formal structure
is brought into conformance with XML requirements.  The *meaning* that
can be expressed in markup --- the "document semantics" --- the holy grail
that drives structuralists to religious ecstasy --- is unchanged.

Well... okay, maybe XHTML Basic (<http://www.w3.org/TR/xhtml-basic/>) is
better.  Hmmm... that gets its semantics from "Modularization of XHTML"
(<http://www.w3.org/TR/xhtml-modularization>).  Okay, now we're... nowhere.
I can't find a semantic definition in there.  Lots of syntax; no semantics.

W3C hasn't issued any new markup semantics since HTML 4.01.  There has been
no subsequent movement towards "the semantic web" --- everything since is
semantically equivalent to HTML 4.01 or some subset thereof.

Now I want to repeat, and emphasize, a sentence fragment I wrote earlier:

     The problem of defining the semantics of a particular XML application
     is identical to the problem of defining a markup language standard.

Alas, we don't get something for nothing.  The wish to define a public XML
schema with more useful, powerful or convenient semantics than existing
XHTML poses precisely the same problems as would be encountered should W3C
decide to issue "HTML 5."

The switch to XML is for software vendors.  It very well may get us less
expensive, more prompt and more robust application software (which are
hardly trivial accomplishments!); but it has nothing to do with the
usefulness of language to authors, the results experienced by users, the
promotion of "the semantic web" or "the separation of style and content"
(or structure and presentation, or markup and rendering, or whatever).
Randall Joseph Fellmy aka Randy@Coises.com
Received on Monday, 19 August 2002 05:59:28 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:03 UTC