DAIS Telcon 13/05/04 ==================== Chair: Dave Pearson Minutes: Mario Antonioletti Present: Mario Antonioletti, EPCC Brian Collins, IBM Dave Pearson, Oracle Vijay Dialany, University Southampton Amy Krause, EPCC Simon Laws, IBM Steve Langella, Ohio State Susan Malaika, IBM Agenda: Actions from the previous telcon CIM/DMTF and DAIS (Susan Malaika) Relational and XML Specification Issues (Brian Collins/Amy Krause) AoB New Actions: ------------ [Steve] Send Amy some words about removing document from a collection by using an XPath expression for possible inclusion in the XML realisation. Actions Resulting from previous Meetings: ----------------------------------------- [Dave] Email Greg and Malcolm to determine a time when they might be able to join a telcon in order to see how to take the composite request framework forward. Dave will send a reminder. [Simon] Document the DAIS folks that are liaising with other GGF WGs. Dead Actions ------------ [Norman] Talk to Allen to suggest having a discussion of the OGSA Data Service document at GGF12 and check to see whether Allen wants to have a session at GGF11 to discuss the Data Services. Done. ----- o CIM/DMTF and DAIS (Susan Malaika) Susan: wrote a scenarios doc for GGF10 ... The phone link that Susan was using was breaking up so she went off the call to find a new phone. She did not come back but sent the summary email included below: - Brian Collins and I have been on recent CGS calls. Larry Flon is the GGF CGS (CIM Grid Schema) Working Group Co-Chair. Larry is handling the DMTF Collaboration for DAIS. Perhaps Larry or Andrea (DMTF rep) can join the DAIS call today at 9am pacific - CGS Information is here https://forge.gridforum.org/docman2/ViewCategory.php?group_id=61&category_id - The CGS Charter in support of DAIS is here https://forge.gridforum.org/projects/cgs-wg/document/First_Draft_New_Charter_for_CGS-WG/en/1 The motivation for the collaboration is that DAIS was defining items such as "database capabilities", which the DMTF already models. DMTF CIM (Common Information Mosels) are already supported by many tools. The GGF CGS-WG had already been collaborating with the DMTF. We agreed that we should ensure a common approach for modelling and representing the characteristics of data across DAIS and the DMTF. CGS-WG kindly agreed to help. - A scenarios document was produced for GGF10 to focus the CGS group's extensions for the DMTF models for data and databases. Larry asked for more work on the scenarios document, which I'm looking at. See the scenarios document here: https://forge.gridforum.org/docman2/ViewCategory.php?group_id=61&category_id=626 - Larry has produced a model for columns and tables: See the attachment to: http://www-unix.gridforum.org/mail_archive/dais-wg/2004/05/msg00004.html - The latest minutes are here https://forge.gridforum.org/projects/cgs-wg/document/Minutes_of_Telecon_11-May-2004/en/1 - The next call is on Tuesday May 18th at 9am pacific - We'll be looking more closely at data types and scenarios. We hope you can join us. o Relational and XML Specification Issues (Brian Collins/Amy Krause) Brian sent out a presentation to the mailing list: http://forge.gridforum.org/projects/dais-wg/document/WS-DAI-Relational-Recommendation/en/1 The relational realisation, now called WS-DAIR, has got up to version 0.4. Presently only in slide format. Had a f2f for the editors in London last week. The slide set distributed represents the present view. Slide number 3 presents the changes that have made to the spec. Four major changes are: - Had 3 access operations: query, update and execute. Now only have one - execute. The query and update operations were really mappings to JDBC. Also update and query were just subsets of execute. This fits better with the presentation in the xml realisation. A number of people have expressed an opinion that this would be a better approach in the past. - Had a DBOperationAccess has been renamed to DBResponseAccess aimed at getting the responses from stored procedure. We need something to get at all the different sort of outputs that can come out of execute. - We've never really catered for all of the exceptions that come back from relational databases. This is not standardised. There is SQLCommunicationArea that we can borrow from in JDBC. Can now bring exceptions and warnings from the database back to the client. - Rationalisation of some of the keywords. Renaming for consistency across the other recommendations. Mixed use of capitalisation ... from now on the first character of an operation will always be a capital letter across all the DAIS specs. Inappropriate to go over all the slides now. Have outlined the most important changes. Currently moving the slides into the document format. Have an editors telcon next Monday. Should have a document version of the spec by then. Document submission is on the 21st of May. Short discussion regarding the outcome of the telcon last week where Norman had got approval to submit the DAIS documents on this date. Currently on target to have a first draft over this Friday and will probably iterate through several other drafts next week. Amy on the XML spec. The main changes from the last version were: - Added xml sequence interface similar to the relational rowset to represent results from the XQuery or XPath access interfaces... - Moved xmlcollectionmanagment back to xmlcollectionaccess ... have done this before. This interface allows you to create sub-collections. - The XPathFactory portType has disappeared and an equivalent to the createService operation has been added to the XPathAccess access portType. This interface no has two operations that can either return the result set or a service handle to the result set. Similarly with the XqueryFactory interfaces. - Trying to align with the other specs. Steve: gone through the latest version ... collectionAccess ... should there not be an operation to remove documents using an XPath expression? Amy: can note that down Steve: could write some text... Amy: but one cannot remove parts of a file ... Steve: ... use the XPath on the collection... also XPath query .... should we be able to specify a response format in the query ... if you put an XPath query can you specify that you want back handles instead of the documents? Amy: that's what the factory does ... it has not been removed ...has just moved to a different portType... Steve: ok ... in the xmldocumentfactory it does not take a collection name - should it not take a collection name? Amy: yes, I will have a look at that ... the collection name could be part of the document name ... will have a look at that. Dave: Steve, it would be good if you summarised your points in the email that you send to Amy about using XPath expression to remove documents from a collection. Mario: have taken some notes at the editors f2f ... these are not minutes ... should these be disseminated .. Dave: if it helps with the rationale as to why decisions were taken then they should be aired... Mario: ok will do as Simon suggested and post on gridforge and make people aware of it Dave: can refer people to it if folks question decisions.... so everything seems to be on track Brian, Amy: yes. Brian: have to do several iterations to get the specs in sync in relation to the core document. Amy: will send out a draft version of the xml spec to the mailing list tomorrow for people to comment on. Dave: you can do that with the relational Brian? Brian: yes ... probably ... maybe the weekend though. Susan sent out an email in case she could not get back on the call. Dave summarised this. Brian: to comment Dave, I've been on most of the CIM/DMTF calls and progress has been fairly brisk ... Dave: one thing that we talked about was arranging a meeting with CGS ... has that been talked about ... do not feel that it's really necessary as there seems to be a lot of collaborative work ... ... o AoB Brian: had a telcon on Tuesday with Peter Clarke and Neil Chue Hong for data area chairs for a files BOF... no one else dialled in ... have had some feedback from other groups that, now that the charter has been changed, there is no overlap .. BOF will go ahead at GGF11. Have asked for dial-in access for the BOF for those folks that will not be able to attend. Details will be emailed out nearer the time. Mario: Norman is trying to set up an OO Database realisation telcon probably next week ... time has yet not been finalised. Dave: next week we discuss document issues ... one day before the submission date ... hopefully there will be none ... might be useful to use that telcon to also discuss planning.