Re: Dealing with large collections [Re: URI constraints]

[Danny Ayers <danny.ayers@xxxxxxxxx>]

This is a mighty interesting problem/thread, and I'll get back to
sitting and watching. But one point that came in early on deserves
comment -

[[
Is there any precedent for using XML element names as the content of an
element or attribute? I can imagine a swamp of namespacing issues. Is
there a better way to pick out an element?
]]

XPath? 

Reading between the lines I agree with the sentiment that support for
a full query language ( XQuery, Sparql, something new and marvellously
versatile)  would be overkill, but maybe something like:

SELECT  entries FROM [URI] WHERE [XPath] LIMIT [N]

(Results would always be a set of entries, URI would point to the
collection endpoint, XPath would operate like an XSLT @match and N
would be the max number of entries required).

- would be adequate. 

How this is phrased is another matter - I'm guessing it should be
expressable in WebDAV, as a GET it wouldn't be difficult, alternately
it could easily be wrapped in XML.

Cheers,
Danny.



-- 

http://dannyayers.com