- From: Stefan Decker <stefan@DB.Stanford.EDU>
- Date: Mon, 15 Nov 1999 00:45:44 -0800
- To: www-rdf-interest@w3.org
Hi,
looking at Sergeys Model API, it directly reminds me of an algebra.
XML Query Folks is now developing an algebra for XML, which
aimed be a sound basis for querying XML and for the XML query language
standard.
A little bit more formalized approach would also be useful for
RDF/RDF-Schema, such that we don't have to talk about APIs, but
about operations inside an algebra.
This should be done in parallel with the deployment of APIs.
The three operations on models:
union: Model X Model -> Model
difference : Model X Model -> Model
intersection : Model X Model -> Model
seem very reasonable for a first start of an algebra.
However, these are certainly only the most basic operations.
One objection from my side:
RDF enables distributed Metadata/Knowledge Bases.
However, RDF itself does not include the aspect, that it is
distributed, into account: it is not part of the datamodel.
However, there must be a possibility to deal with information from
unreliable or not trustworthy sites. To do so we have to
dertermine
the source of an RDF statement. Is long as this is not
part of the datamodel itself this information is lost.
An algebra should also include this.
I will collect requirements for the RDF-algebra.
BTW: Compared to the XML-Query-Working Group we are in a much better position.
Querying XML is far more complicated...
CU,
Stefan
Received on Monday, 15 November 1999 03:48:13 UTC