Lists Home |
Date Index |
Rick Jelliffe wrote:
> Daniel Schierbeck wrote:
>> I've read the W3C XML Fragment Interchange Candidate
>> Recommendation, which I've never seen implemented (that may just
>> be because I'm ignorant.) The specification seems verbose, and is
>> therefore in my opinion unfit as a generic way of transferring XML
> I think it has only been implemented twice (once by me, once by a
> student I was helping).
> FragInt is fine technically, but it doesn't really fit in with XML as
> it is practised: if you want a selection of data you use a database
> query that generates XML; if you want a fragment of a book then you
> typically want a chapter sized chunk (file size); if you want to
> encapsulate some fragment you use simple SOAP; people don't do
> distributed updates (or they use backend tools for replication); there
> is no widespread way to say which information in the context has
> inheritable scope, so there is no automatic way to make fragments of
> that kind of documents; etc.
> I would love FragInt to become a standard part of the basic XML stack.
> Its absense is an XLink killer, for example. But probably it is one of
> those things so fundamental that there is no demand, perhaps in a
> similar fashion to how the tribesmen in my brother's old village never
> wanted soft underpants instead of itchy arsegrass: never had it, never
> missed it.
> Rick Jelliffe
That's why I'm proposing a simpler way of doing it: simply encapsulating
the fragment in an <xml:fragment/> element.