XSL Debate, Leventhal responds to Stephen Deach

[Simon St.Laurent]

At 03:05 AM 6/23/99 +1000, Rick Jelliffe wrote:
>Old SGML hacks used semantics to mean any markup not concerned with
>abstract labelling, including formatting. RDF people use semantics to
>mean
>linked to controlled vocabularies.  The usage of "semantics" in neither
>communities seems to match your usage, where semantic markup
>includes "human-guessable" (presumably primarily to those in a dialect
>group).

I think we're going to have to accept that 'semantic' means different
things to different dialect groups, even within the XML community.  I'll
definitely push for semantic to mean meaningful, even at different levels
of meaning.  Automatically machine-readable is not required for my
interpretation of semantic, but human-readable and _potentially_
machine-readable certainly are.  

It'll be a while before we arrive at controlled vocabularies that are
mutually understood - I don't want to see that process hijacked by a
formatting-oriented controlled vocabulary (and its associated processing)
before we have a chance to develop more meaningful vocabularies.

>Limited meaning-semantics are not solved by labelling, but by linking
>to  well-known vocabularies. Like I said, just providing labelled data
>does little (except allow better guesswork, I suppose).

It's an intermediate step that allows us to reach controlled vocabularies
without convening enormous congresses to decide on official vocabularies.
It's not perfect, but it's much better than nothing (aka a formatting-only
vocabulary.)  'Mere' labelling is adequate for a wide variety of tasks.

>Imagine a company that makes money by providing data over the Web;
>the data may be freely available but it is their markup that provides
>the added value on which they build their company. They might easily
>want to provide the public with data in forms that protect their
>labelling and semantic investment.  If they think agents will be
>good for their customers or business, that should be their choice:
>they can generate RDF if they want semantic markup, or just the
>vanilla XML if they want to provide only labelled data.

First, we're not even talking about 'vanilla XML' here - we're talking
primarily about XSL FOs.  Technically, it's XML, but it's scrubbed down to
be without semantic content in the markup.

Is this behavior that the W3C wants to encourage?  I've argued before that
this strikes at the heart of what XML was designed to do in the first
place, and that it's a hazard standing in the way of making the Web more
useful rather than merely more readable.  Treating semantics as 'added
value' is perhaps good for windfall profits, but lousy for the Web and its
development.

If companies are concerned about information reuse, they should look into
other mechanisms for limiting redistribution, including the simple step of
putting a group of copyright lawyers on retainer.  There are other ways to
solve these problems that don't involve keeping the Web as dumb as possible.


Simon St.Laurent
XML: A Primer / Building XML Applications
Inside XML DTDs: Scientific and Technical (July)
Sharing Bandwidth / Cookies
http://www.simonstl.com

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)