Hi Danny --


At 10:29 AM 11/9/2005 +0100, you wrote:
I was wondering if anyone had anything for conveniently manually
managing relatively large amounts of data in an RDF store. Something
comparable to the GUI tools available for most RDBMSs.

One way is to store RDF as triples in a vanilla SQL database, then use a tools such as our Internet Business Logic system to query and update them.

The example www.reengineeringllc.com/demo_agents/RDFQueryLangComparison1.agent illustrates this without SQL (flat file of triples), while www.reengineeringllc.com/Oil_Industry_Supply_Chain_by_Kowalski_and_Walker.pdf
shows how this extends to SQL.

HTH,  -- Adrian


INTERNET BUSINESS LOGIC (R)
Online at www.reengineeringllc.com
Shared, community use is FREE.

Adrian Walker
Reengineering
PO Box 1412
Bristol
CT 06011-1412 USA

Phone: USA 860 583 9677
Cell:    USA  860 830 2085
Fax:    USA  860 314 1029



At 10:29 AM 11/9/2005 +0100, you wrote:

I was wondering if anyone had anything for conveniently manually
managing relatively large amounts of data in an RDF store. Something
comparable to the GUI tools available for most RDBMSs.

What I would ideally like is something that could do some
(table-based?) visualization with filtering, so unwanted triples could
be sifted and deleted  (I'm not so interested in the DL aspects, just
basic instance data management).

I realise there are tools available (Protege etc) that offer a lot of
manipulation facilities, but I think all those I've seen would require
dumping the data from the existing store, loading that into the tool,
processing as needed then pushing it back again. But this seems a lot
of unnecessary work, surely it's possible to address an existing store
directly, more generically? This strikes me as being something
generally desirable though obviously needing quite a bit of work to
develop.

A possible approach that comes to mind is having a GUI generate SPARQL
queries on the fly to populate a local model from the live store, then
the user manipulation being translated back into SPARQL+delete/update
which goes back to the store. I don't know how feasible such an
interface would be on existing tools, but it seems reasonable in
principle at least...

Another possible approach, again I'm not sure how sane, would be to do
query rewriting the other way than usual - taking SQL queries and
translating them into SPARQL (+delete/update). If that were feasible,
then the SQL views could maybe be addressed through an existing RDMBS
GUI tool.

Failing a general solution - anyone got a good tool for Redland?

Cheers,
Danny.

--

http://dannyayers.com