GGF NM WG Notes 8/31/04 Call began at 11:00 am & ended at 11:34 EDT. Attendees: Martin Swany (UDel) Dan Gunter (LBL) Loukik Kudarimoti (DANTE) Paul Mealor (UCL) Russ Hobby (Internet2) George Brett (Internet2) Jeff Boote (Internet2) Eric Boyd (Internet2) Susan Evett (Internet2 - scribe) Eric started the meeting with an agenda that included presentations in Belgium and a review of Items 7-9 in the schema. Eric wanted to come up with a formal agenda for the Tuesday 21 September meeting in Belgium. Mark Leese contacted Eric to suggest that the group not plan to meet until at least 11 am so that members of the group could attend a conflicting meeting that is scheduled for earlier that morning. The group did not have a strong opinion. Eric asked Dan what talks he was planning. Dan reported that he thought Martin was going to give a talk on revised Response schema; Martin believed he would need about 20 minutes to present and allow 10 minutes for discussion. Eric was going to give a talk for Warren on the implementation of the Request schema in piPEs (again 20 minutes plus discussion). Eric asked Dan if he could give a talk on reconciling the Request and Response schemas. Paul was requested to give a talk with Mark on the revised Request schema (Eric allowed 60 minutes for that due to the length of the material). Eric asked for other suggested topics: - Dan suggested time stamp note – 30 minutes including time for discussion. Might also include 2-4 minute discussion on name space – how we’re going to name at the top level for all the projects. - Dan will give a demo of the Internet2-Netlogger database; this will evolve into a demo of the new schema that Martin is going to talk about. Eric noted that Martin’s talk would be followed by this one. Dan estimated that he’d need 15-20 minutes, max. - Loukik would like a discussion of verification of schema (Loukik and Paul) – they can show a demo that the schemas are working and then a discussion of how it is working. They have verified that requests are coming through. Again, 30 minutes, maximum is all that is needed. - (From later in the meeting chronologically): Business logic was suggested as a topic for the Belgium agenda. Eric defined it as the legal responses (including error handling) to requests. How do you write the program that generates the response? (i.e., you have an optional field with a default value, how do you represent that?) Martin suggested that the group might not have sufficient experience to implement this; Eric felt that the Advisor and MonALISA teams are already working in this area and the group should give them some direction. - Dan wanted to talk about how to handle errors before the group goes into business logic. ACTION ITEM: Eric will send out a tentative agenda this week so for confirmation, additions, and changes. Eric asked if Paul could lead the discussion of Items 7, 8, & 9 or if the group should wait for Mark’s input. Paul believed the group should hold off discussing items 7, 8, & 9 until Mark could attend. Eric suggested that the group briefly review. Item 7, 8/10/04 Item 1C – group has discussed whether you need parameter tags in both locations or if you could use tags for request and something else for response; Eric and Paul agreed that this would be difficult to implement so it would be better to suggest just using tags. Item 8: Paul recalled that this was a discussion on what parameters should be added and how one would go about adding items. Eric suggested that the group name x number of items and, if a user wants to add more, they name it themselves. Dan felt this was a schema validation question – need to be careful about adding items. Eric asked if they could have a “custom” tag – the user would call each new item “custom_itemname1” and “custome_itemname2”. The schema could recognize any “custom_x” tag in the same manner. Dan’s concern is that it would start generating stubs and the system wouldn’t know what to do with them. Eric asked what the best thing to do was; Dan said he’d never seen a system that works so the best method is to stop validating the schema. Dan says that the validation is tricky; the group might want to look at what the WSRF is recommending. Eric asked if anyone was willing to follow up on this – no response. ACTION ITEM: Someone in WG will review what WSRF is recommending and report back to the group. Dan felt Martin is right that this is a tough issue – Eric suggested that the group table this discussion and have Martin address it as part of the revised Request schema. Item 9 – Eric noted that, right now, we don’t have a ‘state’ machine; a request comes in and a ‘state’ is not needed. However, in the future, this might be needed and, for debugging purposes, might be desirable. Dan commented that this would be necessary for multiple requests and partial responses; Eric suggested that client could self-identify w/name and number and the responder parrots back that info. Martin agreed that could work but also suggested that a client makes a request and gets a server-generated ticket number. Eric suggested ending early and will talk to Mark to see if he can meet with the group either next week or in two weeks; the next meeting will be scheduled such that Mark can participate. ACTION ITEM: Eric will send out invitations/reminders when the date is set. Call ended at 11:34.