[
Lists Home |
Date Index |
Thread Index
]
It comes down to naming the reference for the semantic authority:
1. DOCTYPEs can do this with a PUBLIC identifiers but the
scope is the entity.
2. Namespace URIs can do this if you can decide what is on
the other end of the locator.
3. Comments can do this but might be stripped.
4. PIs can do this but might not be passed.
5. ANYUri in an attribute can do this but you have the
same problem of item 2 and you have to fix a value for
the att name so you are back in architectural forms land.
This gets even weirder with aggregate documents (aka,
mixed namespaces). The emails describing UI/nonUI issues
on the TAG are deja vu for the MID project (you can't
crack the nut of hypertext documents without admitting
or obscuring the separation or union of controls with
content, so you will produce something very abstract
and not very useful, or something very concrete and
not very evolvable and wind up back in architectural
forms or class hierarchies for networked components).
Or you punt this stuff away and weaken the warranty
by loosening the fit and function. I tend to agree with
Tim Bray's comment here, at the Berkman meeting and
elsewhere that we do better to talk about the data
we pass around and leave the driving to Hertz or Avis.
As a vet, my worst days were trying to make a one size
fits all specification for hypermedia systems. It
wasn't that it couldn't be done, but that it had been
done repeatedly and it looked like Windows every time
until I worked with real-time 3D. The expansion there
was to get away from buttons and menus and to model
real objects interacting in real time via a network
signified by routed events.
For real time systems, the organizing principle is the event.
The rest is mostly plumbing and layout. Now figure
out the difference between a message and an event.
Just try... :-)
len
From: Chiusano Joseph [mailto:chiusano_joseph@bah.com]
> No, the tags in your markup are arbitrary strings. It's your
> description of the markup "When I say Name, I mean personal
> name" that conveys the semantics. There's arguably a default
> description "When I use a tag that's an English word, I use
> it with the same meaning as that English word", but that's
> still something that's external to the XML document itself.
> The tags have no meaning without an external explanation of
> their intent and usage.
Not sure if anyone is doing this (or has thought about it), but it seems
to me that if one included a reference (URI) to an instance of an OWL
class (or property of a class) then that can be a step in the right
direction.
So perhaps one could have:
<Name owl-ref="www.someuri.com">Peter Smith</Name>
where "owl-ref" (attribute name simply chosen by me for this response)
refers to an OWL class/property that conveys that "This is the full name
of a Person".
|
