- From: Eric Prud'hommeaux <eric@w3.org>
- Date: Thu, 1 Apr 2004 13:54:10 -0500
- To: public-rdf-dawg@w3.org
While we're contemplating negation (safe, stratified, mystified), we should also deal with the practical problem of aggregate functions. Find me the cheapest flight (in the database) from Boston to Chicago. The merits of the use caes are pretty obvious, it's the sort of problem we solve whenever we select from amongst a finite set of options. It's also pretty contentious as it specifically closes the world -- it causes the server to provide answers that it may retract if it encounters more data. My two cents: The semantics of the query language don't have to follow the semantics of the data it queries. It can allow the user to tacitly assert a premise like "...that you know of right now". People intuitively grasp this constraint and check multiple sources for things like cheap flights. A couple more cents: An agent that is processing a portion of an RDF document (eg FOAF scutterers looking for some FOAF and DC arcs and ingoring the rest) must be confident that the arcs that it *doesn't* interpret don't contradict its interpretation of the arcs it *does* interpret. Hence you don't come along after the fact and change meaning of a foaf:knows arc by adding arc saying "well, doesn't know yet, but will meet soon". That would break the interpretion of foaf:knows that everyone had programmed into their heads and their agents. I don't think we have such goals for queries. A query caching mechanism could, by the same analogy, be hampered by the results to queries changing as the knowlege source gets more information. And certainly the querier will be expected to publish the results reponsibly (and not tell the world that for time, UA836 is the cheapest flight form Boston to Chicago). But I think this use case is common enough that it may be more practical to allow aggregats and stratified negation and make sone constraints about the interpretation of the results. -- -eric office: +81.466.49.1170 W3C, Keio Research Institute at SFC, Shonan Fujisawa Campus, Keio University, 5322 Endo, Fujisawa, Kanagawa 252-8520 JAPAN +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA cell: +1.857.222.5741 (does not work in Asia) (eric@w3.org) Feel free to forward this message to any list for any purpose other than email address distribution.
Received on Thursday, 1 April 2004 13:54:22 UTC