14:58:21 RRSAgent has joined #RDB2RDF 14:58:22 logging to http://www.w3.org/2008/09/19-RDB2RDF-irc 14:58:40 meeting: RDB2RDF 14:58:52 Andre1 has joined #RDB2RDF 14:59:12 chair: Ashok 15:04:51 iv_an_ru_ has joined #RDB2RDF 15:05:48 soeren has joined #rdb2rdf 15:06:08 scribenick: soeren 15:06:26 scribe Soeren Auer 15:07:40 no objections for previous minutes 15:08:01 oct30/31 will be f2f minute organized by wolfgang 15:08:11 wolfgang will send more information 15:08:58 no phone calls for next two weeks and no call 17th and 24th of Oct 15:10:04 jenny green provided interesting use case scenario 15:10:33 orri updated RDB2RDF requirements 15:10:53 This is the link from Orri's updated requirement document: http://esw.w3.org/topic/Rdb2RdfXG/ReqForMappingByOErling 15:12:17 discussion about Orri's requirements document 15:12:42 Orri gives an introduction 15:16:07 better definition of the scope of the requirements document 15:16:23 addition of usages scenarios 15:16:47 esp. with regard to mapping on demand 15:17:26 some kind of basic inferencing is required 15:17:47 e.g. when the same triples can be derived from different sources 15:19:07 properties of mapping functions (injective, bijective etc) must be known by the mapping layer 15:22:30 virtuoso provides some of the functionality already 15:23:21 question from Ashok about Simplicity 15:23:44 How can that be achieved? 15:24:05 Orri answers: maybe by two language layers. 15:25:40 often only few simple things are required 15:27:29 Soeren: why not use SQL as a mapping language? 15:29:09 Ashok: You can define a SQL query for each OWL class 15:30:30 Orri will collect some examples of thnigs which are not easily expressible in SQL. 15:33:57 Jenny: If you limit mapping to SQL you n only solve problems that SQL can solve 15:34:01 Limitation to SQL can only problems that SQL can solve. 15:37:12 Ashok: benefit of translating from RDB to RDF is integration of data, but the real value is when you can add extra semantic 15:38:08 +q 15:38:51 Ivan: SPARQL is easier and simpler to write 15:39:23 Jenny presents her use case: 15:40:45 AxelPolleres has joined #RDB2RDF 15:41:03 create a demonstrator that takes spatial operations into account in RDF 15:41:17 5 different data sources from different companies 15:41:54 questionnaires and knowledge elicitation techniques employed to create the data 15:42:45 creation of ontologies for the datasources 15:42:58 now merging of the ontologies and data sources 15:43:15 meeting: RDB2RDF 15:43:17 spatial attributes hae to be considered 15:44:23 SQUIRREL used for linking 15:44:55 starting point are SPARQL queries 15:47:02 one big table for polygons 15:47:47 seven different conditions define what is a field, many different constraints on the metadata 15:48:45 use of Oracle spatial operations 15:49:31 D2RQ used for mapping 15:52:36 no change of underlying data 15:53:23 sometimes preprocessing required 15:54:01 Orri: interesting use-case 15:54:51 Ashok: Excellent use case since it speaks about data from multiple DBs and allows addition of semantics 15:55:25 one more phone call till F2F 15:55:33 10th of October 15:56:00 continuing discussion of Orri's requirements and Jenny's use case 15:56:27 in the next conf. call 15:56:59 Orri promises additions till next call 15:57:34 Adjourn 15:57:52 rrsagent, make logs public 15:58:10 rrsagent, create minutes 15:58:10 I have made the request to generate http://www.w3.org/2008/09/19-RDB2RDF-minutes.html Ashok 18:23:07 Zakim has left #RDB2RDF 19:18:25 MacTed has joined #RDB2RDF