[
Lists Home |
Date Index |
Thread Index
]
Paul Prescod <paul@prescod.net> wrote:
| Arjun Ray wrote:
|> That multiple taxonomies are simultaneously in scope is ultimately a
|> feature of the instance document.
|
| It can be. It can also be a feature of the document type.
Well, yes, I went on to mention that; but the fact that the feature in
application could be instance-specific means that there *is* a context in
which "inline" markup is necessary.
| i.e. sometimes you are doing "CONCUR" and sometimes you are doing
| "subtyping". I consider it a flaw that architectural forms conflate the
| two concepts.
I'm not sure what you mean here. The part of CONCUR that has teeth is
overlapping element boundaries, surely not the consideration here. As for
distinguishing "is the same as" from "is a kind of", the former is usually
a special case (where the syntactically effective "subtype" adds no new
twists of its own); And I'm not sure why pure aliasing needs to be
distinguished in any special way.
That said, the syntactic means available do not exhaust the possibilities.
For an instance specific application, you have an attribute specification
in a start-tag. If the attribute specification applies to all instances
of the element type, you can represent this in an ATTLIST declaration,
with varying degrees of strictness (#REQUIRED, explicit defaulting, and
#FIXED). What cannot be done conveniently is to allow cancellation of a
default (i.e. to say in the document definition that the association
applies "except where noted otherwise" - which would also call for
instance specific markup, btw.)
But in any case, the point of the markup is to register an association.
What classing or typing or subclassing or subtypinf have to do with this
per se isn't clear. So, could you clarify the flaw you have in mind?
|> It is gratuitously restrictive to require that this be defined only in
|> terms of "document types".
|
| I didn't require that *in general*. I said that *for HTML* it is not
| appropriate to require extra attributes in the document.
Okay, here you're talking about things which can be considered part of the
definition of HTML itself - such as href for linking. You're right, an
HTML instance doesn't need to say "my href is an XLink href" if this fact
is a given in the very definition of HTML.
I couldn't agree more.
|> But you won't get away from an irreducible minimum, which is to declare,
|> somehow, within the instance - that is, er, inline - that such ancillary
|> information is in play.
|
| I disagree. As long as there is any other information in the document
| establishing a globally unique name (e.g. a DTD declaration or namespace
| declaration) I can use that as a key to invoke my XLink recognition
| process.
You just conceded my point, with "any other information". I was claiming
the necessity, not the particular form. (What form of "DTD declaration"
did you have in mind?)
|
