[
Lists Home |
Date Index |
Thread Index
]
Our annotation nodes are elements in a separate document or documents.
These have an href attribute that points to what is being annotated.
The HREF uses a restricted XPointer syntax:
URL#xpath(LocationPath)
Where xpath() is a simplified XPointer scheme that supports just XPath
LocationPath expressions.
Our annotations are very nearly [and designed that way] simple XLinks.
My real issue about context is the selection of documents [linkbases] to use
when searching for third-party links. If I have a given document, and can
get to its linkbases [by getting some metadata stored with the document],
then I can find the annotations that I want.
I'm thinking of just standardizing on XLink in general for annotations.
Then, I can somewhere declare the roles [and arcroles] that I want to
reverse index [possibly collection level metadata]. So then, if I add a
document to a collection, and that document contains links with the given
roles, I would locate the nodes referred to in the @xlink:href of the
element to find the nodes, and reverse index them. If the links fail to
execute, then I fail to add it to the collection, and report an error back
to the caller. Importing a link element should also trigger indexing on the
element and its attributes, in this system, since the identification of a
link in a document implies that there is a high likely hood of its
attributes being needed [at least in the system I'm working on].
Keith
<<attachment: winmail.dat>>
|