- From: Richard Newman <r.newman@reading.ac.uk>
- Date: Sat, 29 Jul 2006 22:29:18 -0700
- To: "Xiaoshu Wang" <wangxiao@musc.edu>
- Cc: =?ISO-8859-1?Q? "'Reto_Bachmann-Gm=FCr'" ?= <reto@gmuer.ch>, "'Semantic Web'" <semantic-web@w3.org>
... and a 2¢ input to the rest of the discussion, too: Accept-vocabulary does not put a burden on the ontology developer. It does run a risk of drastically reducing the amount of information sent in response to a request, but that is not necessarily a bad thing. Ontology developers -- and, indeed, data producers -- should use the correct terms regardless. Assuming that Accept-vocabulary is defined as "I would prefer to receive this vocabulary, but can accept others", then a sufficiently- smart server can try to provide suitable triples as a representation of the resource. A dumb server can return whatever it likes, including other vocabularies, ignoring the special request entirely if it so chooses. Assuming that it is defined as "I can only use vocabularies X, Y, and Z; anything else is useless to me", then a sufficiently-smart server can use whatever means necessary to provide an RDF representation about the resource in those vocabularies to the client. If no such triples exist, it can return an empty graph. A dumb server can simply return the subset of triples that it would have returned to a conventional request -- consider this server-side vocabulary filtering. I can imagine the former being very useful for automatic ontology translation; if I maintain a personal profile, maybe linked to a database, I can tell my server a whole bunch of mappings (perhaps to FOAF, vCard or custom ontologies), and allow it to answer requests using the terms that the client wishes to use. I can imagine the latter being very useful for limited devices, such as mobile phones. A phone is unlikely to have a smart reasoner with ontology mappings, so it *really* can't use triples in terms of vocabularies for which it was not designed. Separating the wheat from the chaff before transmission seems helpful. -R On 29 Jul 2006, at 8:52 PM, Xiaoshu Wang wrote: > Problem 1: an RDF document can be written in multiple similar > vocabularies. > Use Accept-vocabulary to ask the server to return the statement > written in > certain vocabularies but not the other. > > My position for this problem (let's call it alternative vocabulary > problem). > It is O.K., fundamentally. But I don't think it is practical. It > puts too > much burdern on ontology developer. It is cheaper and easier doing > this > sort of things at other places. > > Problem 2: Let's call this subgraph problem. I.e., the Accept- > Vocabulary > ask the server to return only those subgraphs that the client request. > > My position to this problem. No. It is fundemantally wrong and we > should do > it with a web service etc., using SPARQL.
Received on Sunday, 30 July 2006 05:29:32 UTC