DAIS WG - OGF20 - 09/05/07 ========================== Chair: Dave Pearson Notes: Mario Antonioletti 13 Attendees Present Presentations in pdf and powerpoint format are available from: http://tinyurl.com/37wcmg Dave Pearson opened the meeting and gives an overview of the present status of DAIS. Isao Kojima presents the work that has been done on the WS-DAI RDF(S) Realization - Querying Specification. David Martin asks whether RDF is XML and whether the XML specification can be used? Isao Kojima: RDF has a structure on top of XML. ... A glossary has been created for the terms used in both specification documents -DAIS RDF(S) spec and the DAIS Ontology Access spec. Inconsistencies in the terminology between the two documents have not been sorted out yet. ... Why do you need stateful services? In the matchmaker case uses the state of the resource to select the appropriate resource. ... Said Mirza goes on to talk about the use case examples for this document. Can map resource properties as RDF data (ontology instance) so the RDF database ca support a registry. Paul Strong, who gave a key note at OGF20, modelled the EGA reference model using RDF. It may be possible to provide this as a real use case. They have 10000s of servers and software - they want to model this to do intelligent decisions for upgrades and the like, they could see if these interfaces can be used. Dave Pearson will try to chase this up as a real commercial based use case. Jena is used as the back-end database though it is not restricted to only using this. ... Concern is expressed in the use case that SPARQLExecute is doing a query and a query and join. The distributed query example has not been published but HP has been doing similar work. There is interest by Unisys for the Large scale RDF distributed storage the use case presented where RDF is used to resolve meta created by different sensor technologies across the supply chain. It is noted that HP has developed a processing engine for SPARQL. ... It is noted that the specs occupy distinct terminology spaces but still need to be consistent. ... --- Oscar Munoz presents an update on the progress made on the Ontology Specification. A question as to whether the access operations count has gone up or down since the last OGF - not sure. Question about the granularity of the operations - too much fine grained access may require a lot of interactions and overhead. Can use an iterator model? ... Oscar goes on to talk about the Glossary document. ... Dave Pearson sums up by noting that Isao said there had not been a lot of change? This could be taken as a sign of stability. However there still remains a lot more work to do. There is only a small group of people undertaking this hence resources limited. Two disappointing aspects about this: Have solicited feedback but have not had any feedback. W3C said that they should try to do it within that group. The use cases show that there is a demand for this, but there is no adoption. It is noted that Sparql provides its own wsdl interfaces but these do not support the indirect form of access described in the core WS-DAI spec. ... question about how you know the types of an RDF model - this is obtained from the model. The description of the types comes to the clients using different kind of messages sent to the server. ... For sparql you are looking at a defined language - can use sparql to query the repositories but implementors have different ontologies and defined different ways of maintaining and implementing these ontologies - how would the defined operations make use of the existing implementations. OntoGrid (?) have 2 implementations - one for Jena and the other for sesame. Have to do a mapping from the operations to the underlying implementation. ... It would be useful to have rough estimates as to when the specs could be submitted. How much more work does sparql need? Isao: it depends on the response but probably within two or three OGF meetings. On the ontology side - there is funding from another project that could help on this. Need more explanation about the operations and maybe remove some of these. Dave Pearson asks Dave Martin how this work can be promoted within OGF to obtain more use cases and more effort. Suffering from the case that RDF is not well know - probably to come up with a simplified spec that people can use. There is a lot of complexity there that make it more difficult to absorve. W3C is the place. Miguel is at w3C and is publicising this work. Dave Pearson mentions that the Inaugural Europe Semantic Technology is going to be held in Vienna - Dave will reference this work and try to promote the work and get more use cases. Dave will get a use case from Paul Strong. Dave asks for a co-chair to get involved.