W3C home > Mailing lists > Public > www-html@w3.org > April 2000

Re: Custom markup

From: Clover Andrew <aclover@1VALUE.com>
Date: Thu, 13 Apr 2000 16:27:47 +0200
Message-ID: <5F78AA062F6AD311A59000508B4AAF6D092AFD@pcs02>
To: "'www-html@w3.org'" <www-html@w3.org>
Yucca <Jukka.Korpela@hut.fi> wrote:

> I'm afraid the Modern Answer to this is: Use XML! All stand and
> praise XML! We're supposed to forget academic ideas about
> universal structured markup and join the chorus.

Eh? XML is universal structured markup, isn't it? There's no
semantic content, but I'd opine that a universal semantic
markup is not practical; one could never fit every detail of
every kind of markup people wanted in documents without making
the language enormously overcomplicated. A general-purpose
document-centric application of XML is possible, natch.
And people are doing it.

> XML offers absolutely nothing as regards to semantics, i.e. the
> meaning of elements in markup.

Quite so. But isn't the problem simply that is is not possible
to define a language to describe the semantics of markup? You
could define a new XML tag, or a new class of HTML element, and
even describe how it's supposed to look, but to give it any
semantic meaning requires that the UA author was already aware
of it.

(We do have a few good languages for describing semantics. They
have weird names like "English" and "Finnish". They tend not to
lend themselves to machine processing.)

> If HTML were developed as a structured markup language for the needs
> of universal hypertext especially on the WWW, then some of the cases
> you mention (like people's names [...]

Well HTML 3.0 had <person>. ;-)

The nice thing about XML is that documents are easily transformable
into languages which do have semantics. You can provide the XML
document for anyone who can understand the semantics of your own
application of XML, and an XHTML version for the average bub with
a web browser. The emboldened name changes from <person> to
<em class="person">, with suitable styling.

Between the two formats - content-rich XML where the semantics may
not be understood, and less-content-rich XHTML (or DocBook, or
whatever) with separate applications that understand their own
distinct semantics - aren't all the bases covered?

-- 
Andrew Clover
Technical Support
1VALUE.com AG
Received on Thursday, 13 April 2000 10:31:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:43 GMT