[nm-wg] summary of 17th November 2004 phone call

[Leese, MJ (Mark)]

Hey folks,

Notes from last week's conference call. Please let me know if I've made any mistakes. Thanks to Susan for taking down the actions, and to those who could attend the call.

Three point summary:
* Dan and Martin's new schema work is progressing well. They hope to circulate a sort of developer's document before the next call, ready for discussion/review on that call. I'm actually reading a draft now!
* We will continue stabilising the current schemas, to be hopefully completed by Christmas. The request schema is almost done. Paul and I will work with Dan on the publication schema.
* Eric will circulate two suggestions before the next call for the next face-to-face meeting. As discussed in Brussels, these will in the US early in the new year.

Actions from the meeting are at the very end of this email.

########## The next call will be Tuesday 7th December (NOTE: not 30th November as orignally planned). Eric or I will send out a phone number nearer the time. The main topic for that call will be the new schema work. ##########

Have a good week ;-)
Cheers,

Mark.




NMWG Conference Call: Tuesday 17th November 2004, 16:00-16:55 GMT
-----------------------------------------------------------------
Present: Mark Leese (Daresbury Laboratory); Eric Boyd, Jeff Boote, Susan Evett and Jessie Goger (Internet2); Dan Gunter and Brian Tierney (LBL); Tanya Brethour (NSCA); Paul Mealor (UCL); Martin Swany (Uni of Delaware).


Agenda:

1. follow-up questions for Dan and Martin on the new schemas 
2. further discussion of the major topics raised during the last 3-4 weeks, i.e. stabilising the current schemas, and inclusion of transaction IDs
3. discussion of outstanding issues to allow stabilization of the current schemas by Christmas

Also discussed:
4. next face-to-face meeting


Contents:

1. follow-up questions for Dan and Martin on the new schemas:

Martin and Dan gave a quick review on the new schemas. They sent out some example instances (ping, iperf and traceroute) on 25th October. Work has continued since then, so we need some new examples (although these could go in the developer's document....see later)

They still need to look at extensibility, and validation via namespaces - tying up parameter blocks to certain requests.

New version provides more flexibility and extensibility  - meets all the requirements document between request and response schemas via commonality between the parameter blocks. 

Version numbers will be needed to indicate each release.

Martin: The schemas will allow incremental additions of "event types". We need the group to vett what we produce, and to come up with new event types. 

Dan is working on a developer's guide, to add explantory text to the examples. He has a rough first draft now - expects to have something sent out by the next call so people can review it and make comments. Martin will write some text for this which will summarise how they see our NM-WG definitions being used in practice, e.g. a collection of WSDL.

Martin: We also need to look at the statistics side of things. It would be impossible to enumerate all of the possible transformations you could do. We think we could prove that they haven't all been defined by the IETF.

There was some discussion of how to introduce new items into the Name Space - NMWG would be the gatekeeper to approval, with interested parties making submissions as and when. There was also some talk of a cookbbook standards document. Users would still be able to define their own items but they wouldn't necessarily work with the other standards. 

Next steps:
i.   produce the developer's documentation
ii.  get feedback from the group
iii. implementations!

In summary, the work has lots of promise, and everything will hopefully become much clearer when we've all had a chance to read the developer's document.




2. further discussion of the major topics raised during the last 3-4 weeks, i.e. stabilising the current schemas, and inclusion of transaction IDs:

Mark: sent out a summary of the minimum set of issues we'd need to resolve on 20th October, and then a summary of the comments he received in response on the 22nd. The two big issues are:
* whether we actually bother stabilising the current schemas or just wait for the new ones - most people seemed happy with this
* inclusion of transaction IDs - this wasn't so clear

Martin: Richard mentioned at SuperComputing that Geoffery Fox had suggested we make better use of available tooling, e.g. leverage WSRF for transaction IDs, etc.

Cue lots of discussion:
* Concern was expressed that this would mean learning WSRF, which might change in the future since there are a lot of early adopters who are introducing changes.
* Paul can show us a way in which a transaction ID can be sent in the same SOAP message as a request/response without being part of the schemas at all.

The group eventually decided that transaction IDs will not be included in these current schemas:
+ most people won't use them
+ Paul has a method to include them in SOAP messages and keep them out of the schema
+ we can have this debate again when it comes to the new schemas

Dan: saw an email from Loukik that said new schemas would have to be backwards compatible with the current ones. This would be so restrictive that it's just not going to happen.
Mark: That's okay, but as already discussed we need to keep people updated about the new work.

So, summary of point 2: Stabilise current schemas by Christmas. Leave transaction IDs till later. No enforced backwards compatiility with new work.




3. outstanding issues for stabilising current schemas by Christmas:

Paul and Mark have worked on the request schema, and are almost done. They think they can work offline on the response/report/publication schema with Dan, saving everyone else from boring calls ;-)




4. next face-to-face meeting:

Eric: When are we next going to meet face-to-face? 
Mark: The next GGF is Korea in March. People thought in Brussels that this was too far too travel, but more significantly that we'd probably want to meet sooner than March. It had been suggested that we could meet somewhere on the US East Coast in January, or tag on to another event like Joint Techs.
Eric: We could meet at Internet2's new offices in January, or tag something onto the next Joint Techs, 13-17th February, in Salt Lake City (http://jointtechs.ornl.gov/SLC2005.html).

Mark: To make it useful we need maximum attendance, so we should probably poll to see which option is best. Either way, we can have more discussion during the next call, when we'll know more about the new developments and whether an earlier meeting (than March) is needed.

Eric will send out an email of possible dates and locations, that states what we'd like to achieve at the meeting.

Call ended at 11:55 am GMT.


Actions:
--------

Dan and Martin: finish developer's doc, including up-to-date example instances (e.g. ping, iperf and traceroute). In addition Martin will add some text which summarises how they see our NM-WG definitions being used in practice, e.g. a collection of WSDL.

Mark, Paul, Dan: Stabilise request and response schemas by Christmas. These can be published via GridForge announced by email. Any announcement should say that new schemas are in development.

Eric: send out an email of possible dates and locations, that states what we'd like to achieve at the meeting.

EVERYONE: review the developer's doc when it's emailed out, and bring comments to the next conference call.
At a later date, we'll be looking for suckers, I mean technically patriotic volunteers to be early adopters, testing out the schemas.