W3C home > Mailing lists > Public > public-xg-rdb2rdf@w3.org > November 2008


From: Juan Sequeda <juanfederico@gmail.com>
Date: Fri, 7 Nov 2008 14:21:17 -0600
Message-ID: <f914914c0811071221la6ba0aaka5466a7acfa76a5@mail.gmail.com>
To: "Sören Auer" <auer@informatik.uni-leipzig.de>
Cc: ashok.malhotra@oracle.com, "Axel Polleres" <axel.polleres@deri.org>, "public-xg-rdb2rdf@w3.org" <public-xg-rdb2rdf@w3.org>
I also agree with Soeren.

In our work[1], where we formalized all the direct mapping approaches (RDB
to OWL), we used FOL to represent the transformation rules. Theses could
then be used in Datalog, or any other rule engine to automatically transform
the RDB to OWL.

[1] http://www.springerlink.com/content/mv58805364k31734/
On Fri, Nov 7, 2008 at 2:16 PM, Sören Auer

> ashok malhotra wrote:
>> The proposal is that RIF be one possible syntax for the mapping language,
>> not the only syntax, correct?
>> Seems reasonable.  A rule consists of 2 parts and in our case the parts
>> may be, say, the name of an OWL class and the SQL query that can be used to
>> query that class,
> I see the use of RIF more in the spirit of using datalog (which are
> essentially also rules) for data integration. So the head of the rule
> defines the resulting class, while the tail selects predicates (from the
> relational DB).
> This seems to be consistent with slide 5 of Axel's talk, which we
> unfortunately did not manage to discuss today since some of the syntax there
> is not completely obvious to me.
> Have a nice weekend!
> Sören
Received on Friday, 7 November 2008 20:21:55 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:39:03 UTC