Re: Minutes from 16th November WCAG WG telecon

At 9:00 PM +0000 11/20/00, Sean B. Palmer wrote:
>No I'm not. Have you seen the list of examples? Have you seen this excellent
>point:-
>[[[
>If you look at the RDF data models, you'll see the reason: if the Semantic
>Web is going to be a machine processable Web of languages relying on
>triplets to eventually lead to proof validation, attaching behaviours to any
>form of semantic processable data on the client side is going to cause
>problems. The schema layer in the RDF model isn't set up to provide
>behavioural display assertion; simply because that task is impossible. By
>attaching behaviours to Semantics that are expected to be recognized and
>displayed properly by default is wrong, and you need to back up the
>proprietary assertion you are trying to make with some type of schema. At
>the moment there is no form of schema available to do this, which is why we
>have Cascading Style Sheet languages to help us cope in the meantime.
>]]]
>Argue with that pure fact if you're going to argue with anything.

That's not really a "pure fact" -- it's just the holy scripture of
what you happen to believe.  (In any case, the whole sentence from
"if the Semantic Web..." to "...is going to cause problems" is
very opaque even to a technical reader like myself.)

>Frankly, the reasons *are* there, but I have no idea about how to assert
>high level principles of Web architecture to someone who asserts that HTML
>isn't SGML, and then goes on to call SGML "SMTL" - and so it doesn't really
>bother me if this discussion deosn't come to a sane conclusion, it's all
>just a bit of fun to me!

The problem is that you think it matters whether or not HTML is SGML.
Ultimately, it doesn't matter one way or another in the majority of
real world applications; in a practical sense, it's at best an
interesting bit of historical trivia.  A dogmatic insistence on
this being important may lead to driving away people who don't
understand or want to understand SGML, and to confusing people who
want a real world, practical answer about why they should do something
rather than one based on theoretical underpinnings of a language
they've barely heard of and never used.

Note that I don't necessarily disagree with any specific items
you've raised -- I simply think that a better answer is needed than
simply the purely academic response of referring to SGML and
pretending as if that is a satisfactory one.  We differ on how
the argument is and should be presented.

--Kynn
-- 
Kynn Bartlett <kynn@idyllmtn.com>
http://www.kynn.com/

Received on Monday, 20 November 2000 16:42:56 UTC