11 Replies Latest reply on Aug 11, 2008 7:18 PM by alwu-Oracle

    What Next?


      Wanted to start this thread to extend this a little further. It is very nice to see OWL support, bulk load api, and other options in 11g that were not supported in 10g. I guess a copule of days back Jena adaptor is also made available.

      I am just curious to see what else is going be added to 11g from semantic web perspective? It will be good, may be if you can give some idea what we can expect in the next 6 months to 1 year.

      [For example I like the way application express team gives the information about their statement of direction http://www.oracle.com/technology/products/database/application_express/apex_sod.html ]

      Thnaks in advance for the input on this question.
        • 1. Re: What Next?
          I apologize for the delayed response to this query.

          It is not customary for Oracle to provide detailed statements of direction. However, here are a few things we are planning for the future:

          (1) Better integration with complementary 3rd party products
          ---eg, visualization of RDF graphs for mining and analysis,
          ontology editors, query tools, etc.
          (2) RDF views on Relational databases
          (3) We are participating in the W3C OWL1.1 group
          (4) Improved performance for loading, inferencing, semantic queries

          Do you have any features you would like to see?

          • 2. What Next?
            Sorry. I did not check your reply until now!! I did check for a response for a couple of weeks. Thanks for the response though. I will be doing some work using semantics web technologies in the bioinformatics area in the days ahead. I will get in touch with you and others in the forum if required. Thanks.
            • 3. Re: What Next?
              I'm working within the area of bioinformatics as well and need very much full SPARQL support (at least the OPTIONAL clause - as mentioned in earlier posts; can't do much without that).

              Now that more that 6 months have gone since the answer from Melli - maybe we could hear a little more?

              bw - Christian
              • 4. Re: What Next?
                With that many views (>1500) it seems like people are interested - couldn't we get some "hints"?

                Regards - Christian
                • 5. Re: What Next?
                  Hi Christian,

                  Full, native, SPARQL support is in the plan. And yes, we have recognized the need to provide at least OPTIONAL, and we plan to release this first. The release date is expected to be soon. I will post updates as soon as we have a concrete date.

                  • 6. Re: What Next?
                    Hi Melli

                    Thank's is sounds good - and of cause full RDFS/OWL support is also very much wanted.

                    By the way, now there is also an "Oracle SQL Developer Statement of Direction" here: http://tinyurl.com/69s342 - so it's not that uncommon for the company to give this sort of info.

                    bw - Christian
                    • 7. Re: What Next?
                      Just to clarify, full RDFS support has been there since 10gR2 :)

                      Oracle 11g OWL inference supports a very rich subset of OWL semantics. It is
                      good enough for many applications.

                      It is also feasible to integrate, using Oracle Jena Adaptor, external complete DL reasoners, Pellet for example, with Oracle 11g OWL inference to achieve more semantics.
                      • 8. Re: What Next?
                        I might be wrong, but doesn't f.ex. rdfs:range = <date> imply that the system should ensure that the property value is actually a <date>?
                        • 9. Re: What Next?
                          The rule for rdfs:range is well defined by the RDFS semantics specification. I am not sure what exactly you mean by "system should ensure ..."

                          Please clarify.
                          • 10. Re: What Next?
                            I think that when a property (f.ex. startDate) has the RDFS:range <date>, the database should give [at least] a warning if one tries to insert a not-valid date as property value (like in RDBMS tables).

                            • 11. Re: What Next?
                              Now I see what you are getting at.

                              Oracle inference is a separate process from data loading. In other words, there will be no inference going on to perform such a check as you want.

                              However, it might be straightforward to write a query that finds such rdfs:range triples
                              and list out non-valid date formats that associate with those triples.