Lists Home |
Date Index |
What you say is both true and wise. However, it seems to me that the specific
context in which Bill Lindsey asked the question--identifying the type of a
document to a native XML database--requires a single specific answer. The
database may, of course, associate many schemas or document type definitions (in
a larger sense than DTDs themselves) with a single identified document type, but
as a practical matter there must be that single anchor for those various
document definitions, or aliases, or as a single target to which to point
different processing strategies. The 'markup' answer to the original question is
that the type of a document is the GI of its root element. The wisdom of that
definition, as expressed by DOCTYPE, seems to become ever clearer as this thread
of discussion progresses. I would argue that, if we allow markup to control,
there might in a native XML database be many document instances which must be
considered of the same type though they exhibit vastly different structures
below the name of their root element. And of course there many be many schemata
and document processing strategies anchored to that one lexically-identified
root element type.
Michael Kay wrote:
> There is no single notion of document type. You can define a number of
> classification criteria that you apply to documents, based on the namespaces
> they use, the schemas they conform to, the name of the top-level element,
> and so on, and you can call any one of these the "document type" if that
> makes sense in your application area.
> Remember that in the real world, types are rarely mutually exclusive.
> (History and Biography overlap, a fact which most bookshops choose to
> overlook.) You can create a document that is both a valid XHTML document and
> a valid XSLT stylesheet.
> Mike Kay