DAIS Telcon 29/06/04 ==================== Chair: Norman Paton Minutes: Mario Antonioletti Present: Mario Antonioletti, EPCC Greg Riccardi, Florida State University Susan Malaika, IBM Norman Paton, University of Manchester Simon Laws, IBM Brian Collins, IBM Dave Pearson, Oracle Amy Krause, EPCC Agenda: 1. Agree process for Socialising standards documents. 2. Agree process for logging and addressing issues in standards documents. 3. Agree process for finalising the mapping scenarios document produced for GGF11. 4. Agree process for focusing on open issues relating to mappings of DAIS specifications. 5. Discuss OGSA Data Services Document and relationship to OGSA Data Architecture activity. 6. Discuss how to proceed with Object Data Access Realisation proposal in time for GGF12. New Actions: [Dave] Post a straw man executive summary and set of questions before the next call for people to comment on. [Mario] Post a summary of existing DAIS issues to the list and let folks know that the Grid Forge trackers are being used for this. [Mario] Propose a process for closing issues that are posted on Grid Forge. Check out what other GGF groups are doing in this space. To be done by the next call. [Norman] Email the people involved in the Object Data Access Realisation to see if this effort can be picked up again. ---- 1. Agree process for Socialising standards documents. Norman starts meeting. Socialising of the standard documents. Dave: this came out of the last GGF - it is now time to go out to the wider community to solicit feedback. Can go to industry (database vendors), e-Science/Research community and the commercial community. Need to be sure to know what we want to find out, what type of feedback we want to get ... an email questionnaire is not seen as being sufficient. Need to entice people to want to do this. Write an executive summary to explain motivation, background and current status ... what needs the specs are going to be meeting ... the key thing is to identify the people to send the questionnaire to. What do people think and what should we ask? Norman: thing we would like to get out of this: more people participating - for them to sign up and feedback and also to engage communities where as yet we do not have much contact, e.g. the xml vendor community ... has anything happened in this space yet? Susan/Amy: No. Were thinking of some groups we could contact ... Norman: maybe should wait for the executive summary before moving on this. ... what input should we be seeking? Dave: we can ask about the general approach of the specification and the relevance - do we have the right interfaces? Can the capabilities of your product be exposed using these method? ... need to pool questions ... need to identify any short comings. We should only put about 10 questions maximum. Simon: it depends on the context in which we ask the questions; whether DAIS is a mechanism of accessing data in the Grid or whether we present it as a general way of accessing data... Dave: the executive summary should make the context clear ... people should still have the freedom to answer what they want ... Norman: may pitch this as being a Web service orientated interface as opposed to a Grid service interface ... Brian: want to solicit more than a yes/no type question and try to elicit constructive criticism or positive endorsement ... there could be a straw man in the executive summary ... Dave: I think that would be the next stage Brian ... need to go through an internal review to see how people might interpret the questions. Norman: if people have specific views of what should be in they should get in touch with Dave. We would like this process to be done before the next GGF. Dave: would like people to identify communities ... Mario: are we going to ask about the thorny issue of mappings? Norman: for this questionnaire we should be gloss over it ... Dave: I don't think we can gloss over it ... some of the history can be given in the executive summary ... I can put something in on mappings and people can comment about the wording ... Norman: at the top level ok ... but not detailed questions ... Dave: it's useful to get a view ... Norman: the questions are not about the nature of the mappings but of peoples' plans and their expectations ... Dave: functionality, details ... I'll pull something together and circulate something by the middle of next week ... --- 2. Agree process for logging and addressing issues in standards documents. Norman: Tracking issues .. ask Mario to comment on this. Mario: currently Grid Forge is being used for this. David Vyvyan put in a lot of issues for the relational spec up on Grid Forge. Norman: ok so Grid Forge is being used ... should make people aware on the list that this is being used. Mario: I will post a summary of issues to the list ... Brian: this is one of stage of the process - have logged the issues and people can post comments but what is the process for resolving the issues? Norman: comments should drift back to the spec team ... now that we have a more public logging process we should have a more formal process. When the spec team makes decisions it post the nature of the decision. Brian: maybe a two step process is best ... put some comments .. Mario: in my view I think the spec authors should be responsible for closing issues - they should post a comment on Grid Forge and possibly also the mailing list. If someone feels the issue has not been properly addressed they can always post a comment and the issue can be re-opened. ... Norman: Mario - can you suggest what the mechanism for establishing process .... GGF does not have a formal process ... see what other groups are doing ... do this by the next call... Susan: the OGSA group uses Grid Forge quite a bit ... Simon: other groups deal with issues in the calls ... Norman: Mario will post --- 3. Agree process for finalising the mapping scenarios document produced for GGF11. Need to agree a process for finishing off the mappings document. Invite Simon to comment on this ... Simon: have discussed this with Savas and Sastry ... keen to add other scenarios ... will have a call to see if this is practical ... to a certain extent there are diminishing returns .... need to do something to bring to this work to a conclusion ... need to do a bit more work but not a lot more. Mario: at GGF people wanted scenarios that brought out the functional differences between the different mappings in the scenarios ... Savas said he had one ... Simon: Malcolm brought up another scenario this morning that did this ... will pool these together to make sure they bring out functional differences and go from there ... --- 4. Agree process for focusing on open issues relating to mappings of DAIS specifications. Simon: problem is relatively fundamental ... disagreement is about the model that sits behind WSRF that differentiates it from WS-I - do we then recommend different mappings for DAIS or do we push for changes to be made to WSRF that would remove these differences? ... Norman: at the last GGF meeting there was one scenario of relevance to DAIS that brought out differences between the different mappings ... singleton data resource ... the thought was that there might be an opportunity to have a meeting before the next WSRF meeting taking place in the UK where this might be discussed. However it appears that Sastry would not be able to make such a meeting ... Dave: have seen response from Sastry ... will send him an email ... we know that he can't come but it would be good to get some input ... there is another Oracle person that we could bring up to speed and get involved. Norman: there is some evidence that OASIS is going to address ... the singleton resource issue identified at GGF ... was hoping that we could come up with a practical mechanism - the result collection case before the next WSRF meeting ... happy to meet up with the Oracle chap coming to the WSRF meeting but we need to understand our own issues first ... Simon: maybe we can do that by documenting the scenario ... not clear what is meant by the singleton pattern ... folks have their own view of what the scenario is and what the solution is ... Susan: ... the 3rd party delivery that matters most to DAIS in relation to WSRF ... Norman: that is my understanding ... can try to get a common middle ground that IBM and Oracle are both happy with by exploring this pattern .. Susan: ... define what the singleton resource pattern might be ... and see how this fits in with the 3rd party delivery and try to feed this back to the WSRF ... Simon: we take the issues and suggest what the solution is and use that as the exit strategy for the mapping documents ... some of these scenarios are more important to DAIS .... Norman: we are going to try to zoom into options for handling the different mappings that might be used for DAIS ... do not have to do in a closed group of three [the authors of the mapping document] ... needs to be more up-front for the core DAIS issues... Simon: not suggesting that this is for the mapping document ... distill stuff from the mapping documents and present these separately ... a link between the two exercises allows us to identify the conclusion that comes out of mapping scenarios ... take the scenario and propose solutions that you then show to people ... need arguments then to show the advantages ... Dave: when you say "this is how we are going to solve it" - are you proposing a specific mapping? Simon: there may be more than one mapping ... can blur between the two approaches ... we do not want to support two different things though... Dave: given that GGF has nailed its colours to WSRF we need to provide a WSRF mapping ... and then start feeding back to OASIS ... Simon: happy to include WSRF ... question is whether everyone else is happy doing this? Norman: completing the mappings documents while the outcome of item 4 may be a decision and not a mapping ... what we need to do is go through a process and document the options for DAIS ... and also explore whether changes to WSRF would make this any easier ... this may not have converged for GGF12 and the mappings document but not so for this discussion ... Simon: ... but we need to start moving towards making decisions ... this has to be written down and not necessarily be part of the mappings document ... Norman: How do we move forward with item 4? Who takes ownership of this? Have two groups: the spec team and the authors of the mapping document ... Simon: pick an owner but bring it to the mailing list ... Dave: what about Savas? Simon: I think he is happy to be involved ... Dave: the people that looked at the mappings are the best people to take this forward and they can make recommendations that the spec team can adopt ... the spec team would welcome a firm decision ... Norman: can bat the task to the Savas/Simon/Sastry team but I would like to join it for the purposes of item 4 ... Simon: but I think 3 and 4 are the same thing ... I think it should be open. Norman: that is why I asked people who would like to join ... would like to have someone that is agnostic in the solution ... not casting aspersions ... the role has to be changed ... no one else put their head over the parapet there ... token is kept by Simon/Savas/Sastry but I would like to be kept in the loop. ---- 5. Discuss OGSA Data Services Document and relationship to OGSA Data Architecture activity. Need to discuss the OGSA Data Services document in relation to the Data Architecture effort currently taking place ... not sure if Allen has had a chance to speak to some of the other authors ... Susan? Susan: I was cc'ed on his emails ... Norman: the question is whether were the authors of the OGSA Data Services document are happy for it to be rolled back to the OGSA Data Architecture document? Susan: did not see people objecting to this ... can try to close this off ... did we not discuss keeping a version of the document about ...? Norman: these things do stay around ... if people are going to refer to one they should refer to the first one ... the second one was more in a state of flux ... happy to wait until Allen comes back ... it does not have to be closed off with any urgency. --- 6. Discuss how to proceed with Object Data Access Realisation proposal in time for GGF12. None of the relevant people were at the call so Norman took an action to email the relevant people. ---- AOB: Susan: in the CGS calls proposed that might be able to release portions of a UML model for relational databases ... looking at this at the moment ... release a portion that DAIS needs ... would be very useful for everyone. Norman: I think that would be very good. Susan: also been looking at the XML ... may need to be some hand crafting to get xml ... it's XMI type format that you would get by default ... looking for tools ... DAIS being able to use the generate xml might be under question. ... Some discussion was had on telcon numbers to be used. Current number was seen as possibly being too expensive for US callers. Some of the discussion of composite requests - would be good to try and get a community practice document ... maybe this could be based on the ogsa-dai approach if nothing else is forthcoming. Brian: discussing with Neil Chue Hong about the files BOF ... should we report at the next call ... do you want a status report at the next meeting? Norman thought that this would be appropriate. ----