- From: Tim Finin <finin@cs.umbc.edu>
- Date: Tue, 15 Aug 2006 18:52:05 -0400
- To: bob@snee.com
- CC: semantic-web@w3.org, Li Ding <dingli1@cs.umbc.edu>, Lushan Han <lushan1@umbc.edu>
Bob DuCharme wrote:
> Why does Swoogle offer the options of searching ontologies, or ontologies
> and instance data ("document"), but not instance data without ontologies?
> This would make it easier for people interested in instance data to get at it.
Bob --
That's a good observation. Maybe we should add that the the
interface.
You can get the same effect, somewhat clumsily, using the
hasOntoRatio operater, as described in [1]. For example,
this queries:
ns:foaf hasOntoRatio:[0 TO 0]
finds documents that use the foaf namespace and are 100%
data. A document's ontoRatio is the fraction of its triples
that participate in definitions. A document with an
ontoRatio of 0 is 100% data. The text after the 'ns:'
qualifier is a token and ns:foaf should match any document
that has a namespace with the token 'foaf' in it.
in Swoogle 2005, we did the retrieval directly in the
database, so our advanced query option was quite flexible.
For Swoogle 2006, the current system, we are doing it all in
Lucene. The result is much faster, but much less flexible.
For example, in range queries, only integers and string are
allowed and the term is expanded by enumerating all values
in the range. We are using the (default) limit of 1024
terms in a query. So, the following queries won't work:
* find foaf documents with more than 1000 triples
ns:foaf hasCntTriples:[1001 TO 999999]
PROBLERM: too many terms generated from range
* find rdf data that might have a few definitions
ns:foaf hasOntoRatio:[0 TO 0.1]
PROBLEM: non-integer in range
Some interesting searches you *can* do are:
* find DOAP data files
ns:doap hasOntoRatio:[0 TO 0]
* find ontologies using the DOAP namespace
ns:doap hasOntoRatio:[1 TO 1]
* find foaf data files written in N3
ns:foaf hasOntoRatio:[0 TO 0] hasGrammar:n3
[1] http://swoogle.umbc.edu/index.php?option=com_swoogle_manual&manual=search_swd
--
Tim Finin, Computer Science & Electrical Engineering, Univ of Maryland
Baltimore County, 1000 Hilltop Cir, Baltimore MD 21250. finin@umbc.edu
http://umbc.edu/~finin 410-455-3522 fax:-3969 http://ebiquity.umbc.edu
Received on Tuesday, 15 August 2006 22:52:12 UTC