[
Lists Home |
Date Index |
Thread Index
]
I like both proposals. These feel like promising paths forward. As you
note, #2 does more than #1.
My primary concern with proposal #2 is that someone would inevitably
pipe up that those descriptions sound a lot like RDF, and then we'd get
to think about all of this in terms of RDF for eternity, plus there are
all the delightful issues of QName use in attributes and how best to
deal with that... so I guess I'd rather stick with an XSLT transform or
CSS decoration.
Paul Prescod wrote:
> I have two proposals:
>
> Proposal #1:
>
> A new vocabulary which we call the XML Linking Schema language. It
> has two purposes. The first is that it describes how to recognize
> links in an XML document, probably using one of the two W3C pattern
> languages (XPath patterns, CSS selectors, ...). The second is that it
> describes the range of each link types. i.e. that personRef must link
> to a person element and so forth. XML Schema has some weak support
> for this (through keys) but it is really not sufficiently powerful to
> be widely usable.
>
> The procedure for finding a linking schema would be just like those
> for finding a schema. You COULD follow a magic attribute
> ("xml:linking-schema"). You COULD follow a link to a RDDL and find
> the linking schema from there. You COULD "just know" what linking
> schema is associated with the top-level namespace.
>
> Proposal #2:
>
> We could have a more generalized syntax for doing out-of-line
> element/attribute renaming or annotating could be standardized. Once
> again it would have patterns but it would be substantially simpler
> than XSLT in that it would only rename or annotate things but never
> (e.g.) reorder them. I imagine it saying something along the lines
> of:
>
> <annotate match="xhtml:img"> <attribute name="xlink:href"
> select="@src"/> </annotate>
>
> The "finding" procedure would be similar to the one before, but not
> link specific:
>
> <mydoc> <xm:type-mapping from="somenamespace" to="somenamespace"
> href="somefile.xmap">
>
> The processor could either use that OR use RDDL OR use some cached or
> a priori knowledge of the mappings allowed for a vocabulary. ==== The
> first proposal kills two birds (link recognition and validation) with
> one stone. The second proposal kills N birds with one stone.
-------------
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
|
