examples as opportunity (was Re: XHTML & hyper...)

Tim Bray writes:
>> (Possible alternatives for such external linkbases, also already
>> written, include RDF and XML Topic Maps.  Other alternatives include
>> some kind of AF-like approach or a vocabulary created with explicit
>> XHTML resonance.)
>
>OK, if you think external linkbases are worth doing, let's see some 
>examples of what they'd look like using some of these alternatives. 

That sounds like an excellent idea, though it feels like a project worth
spending more time on than a mailing list back-and-forth.  I'm also
quite aware that I'm probably not the person the RDF, XTM, or AF
communities, not to mention the HTML WG, would choose to present their
technologies.

There's supposed to be a Town Hall at XML 2002 on hypertext issues - I
believe the TAG's own Norm Walsh is organizing it.  It might be
interesting to model this Town Hall on last year's Schemas Town Hall,
which had sample documents and gave them to different communities to
model with their own tools.

We could do something similar with these technologies, and I can
certainly do XArc as well.  (I'll see if I can find Gabe, who started
that one.)  The HTML WG might want to consider an HLink entry as well.

I don't think there's a mad rush to sort out external linkbases for HTML
in any event, so a December encounter might be reasonable.  On the other
hand, I don't know what Norm's got organized already and don't know if
two months' time is enough to make this all work.

If that doesn't work out, maybe we can arrange something online.
Failing that, I'll be happy to concoct various other flavors of external
linkbases.


To continue with your other comments:
>Well, suppose I want to add multi-ended links to XHTML.  The kind of 
>design I'd like would be real easy to explain to humans and to write 
>software to recognize, it would allow me to decorate the links with 
>titles and explanations and behavior/traversal suggestions, but those 
>decorations would be optional, it would use subelements because that 
>seems like a better design pattern for "repeating anythings" and is 
>i18n-friendly.  It would come with some (optional) precooked behaviors 
>inspired by the semantics of HTML today.    I'd worry a *lot* that when 
>someone did a "View Source" they'd be able to figure out what they saw. 
>    XLink qualifies on this set of criteria.
>
>For external linkbases, I'd be terribly worried about heading down the
>slippery slope that (in my opinion) claimed HyTime, so I'd be brutally
>minimal, just re-use the existing link vocabulary and have one
>hardwired link type to bootstrap a search for external links.  The
>design would look a lot like XLink.

I don't see anything here that expresses how XLink generically is a good
fit for HTML specifically.  It feels only like a list of features you
want and general dislike of HyTime.  You've already judged XLink
adequate, a decision I think other people would reserve for now.

>> Do any of those factors justify making it the only fit for XHTML?
>
> No, but as I said, from those who aren't convinced, I would like to
> see counter-arguments of the form "multi-ended links aren't
> interesting" or "multi-ended links are interesting but here's a
> better way to do it."

So far as I can tell, multi-ended links are most interesting to a very
few people.  I spent a few days last week answering the classic and
appropriate "why anyone needs or would want hypertext links
that are more complex than a named entity and a way to point at it?
In other words, why all the fuss?"

I think in the end I made a plausible case for their use, but that's as
far as I'll claim.


-------------
Simon St.Laurent - SSL is my TLA
http://simonstl.com may be my URI
http://monasticxml.org may be my ascetic URI
urn:oid:1.3.6.1.4.1.6320 is another possibility altogether

Received on Monday, 7 October 2002 10:42:08 UTC