- From: Jim Whitehead <ejw@cse.ucsc.edu>
- Date: Wed, 8 Oct 2003 17:39:14 -0700
- To: "Liam Quin" <liam@w3.org>
- Cc: "Dan Connolly" <connolly@w3.org>, "Massimo Marchiori" <massimo@w3.org>, "Ralph R. Swick" <swick@w3.org>, <www-archive@w3.org>, "C. M. Sperberg-McQueen" <cmsmcq@acm.org>, "Lisa Dusseault" <lisa@xythos.com>
> I think there are two separate questions here... one is about RDF and > one is about XML. Although RDF can be serialized in XML, the sorts of > queries people are doing today with RDF query languages are over the > abstract graph, not over the XML representation. Agreed. > For XML documents, and resources that present themselves as XML, one > might reasonably consider using XML Query. XQuery certainly would be my choice for XML query language. But, even once you've settled on XQuery, there are a number of protocol issues that would need to be addressed. For example, can clients ask that responses not exceed a particular size? Are query responses resources, with a URI, and how long can that URI be dereferenced? How should query errors be passed back to the client? > It turns out there are lots of fairly complex things one might > reasonably want to ask about the contents of XML Documents, so > perhaps it's not surprising that XML Query is a rather large > specification. I wouldn't expect all DAV implementations to support > it, even in XML environments. True, although there is a growing class of DAV servers that sit on top of XML databases, and so are interested in this capability. > So maybe this is an area where technical experimentation is needed > before any standardization is considered. Never hurts. I'm seeing some experimentation among DASL implementors already. > We could go to the XML Query Working Group and see if there is consensus > on any such technical work; my personal expectation is that they are too > busy right now to focus on such issues, just because of where they are > in their editing cycle, but that this may well change by January. This timeframe is probably right on. I think if things got rolling early to middle of next year, that would be fine. Inaction on this does have opportunity costs, but they're still fairly low. - Jim
Received on Wednesday, 8 October 2003 20:43:51 UTC