Difference between revisions of "COSMOS Resource Modeling Meeting Minutes"

Revision as of 12:00, 19 August 2008

Minutes for Resource Modeling meetings.

Dec 5 2006

• There have been several requests for a SML sample. ACTION: S Jerman to ask SML Group when sample is going to be available. [Done - they will investigate]
• Progress report on Validation, IF inport/export: updates added to Wiki page.
• Next meeting 5th Jan.

Jan 5 2007

Attendees: Mark, Valentina, Steve Discussion:

• Progressing on validator. SML workshop may be delayed which will give more time.
• Need to discuss schedule, interaction with other projects, deliverables at Face2Face. ACTION: Steve to add to agenda.
• Discussed need for more samples.
• Next meeting: 19th Jan

26 October 2007

• Working on documenting interfaces and refining issues related to adopters with different data types
• Working on putting together a demo of our CMDBf extensions

David whiteman.us.ibm.com 14:41, 26 October 2007 (EDT)

29 April 2008

Attendees: Rich Vasconi (IBM), Merri Jensen (SAS), Mark McCraw (SAS), Ali Mehregani (IBM), David Whiteman (IBM)

Discussion:

• Gave brief overview of RM to Merri and Mark, members of the SDD team who are wanting to ramp up on SML
• Started to go through open RM bug list in order to identify high priority items for i11
• Ali and David agreed to split up the investigation and assignment of P1 and P2 defects to determine whether they need a different priority, should be closed, and should be i11 or future.
• When the topic of bug 213635 came up, we discussed how the documentation for SML and other COSMOS areas would be delivered via Eclipse. David explained that the current intent is to provide the whole User Guide and Developer Guide as integrated with Eclipse help. Ali suggested that much of those documents, particularly in the UG, would not make sense in the Eclipse environment, since the actors wanting that information would not be using Eclipse. Rich mentioned that he has the ability to exclude parts of the doc from the Eclipse based online help. David said that it would be tricky to decide what to include and how to package it, and that we should look at how other Eclipse projects handle this. This might need to be revisited at a community call or summit.
• We discussed ideas for restructuring our teams on the project. David suggested that if we were not going to be a toplevel project for 1.0, we can now establish the structure of our "workgroups". He mentioned some ideas he had, most notably breaking up the DC team into two, since there is a significant variance in the scope of that subproject. He also mentioned that part of this activity (or all of it) could be in finding people to lead the subteams that are actively involved in the work on those teams.
• Ali expressed concern that more teams would mean more overhead in the project.

13 May 2008

Attendees: Merri Jensen(SAS), Mark McCraw (SAS), Ali Mehregani (IBM), David Whiteman (IBM)

Discussion:

• No new status to report for RM deliverables, other than what is already on the status wiki page
• David explained to Mary and Mark the current focus of the RM team, which is 1.1 spec compliance and bug fixing
• Mark asked about the Tigerstripe integration, and David pointed out that is future work of interest for building SML resources from graphical templates

20 May 2008

Attendees: Ali (IBM), David (IBM), Mark (SAS), Jason (SAS)

Discussion:

• David suggested SDD meeting should become ME meeting, and that it should precede the community call. Jason to send out new meeting time and subject.
• David suggested possibly should do i11 design review, since we have somewhat gotten out of the practice of reviewing enhancement designs
• Everyone feels comfortable with design process described on the cosmos-dev thread
• Current status is up to date on week 3 build status page

27 May 2008

Attendees: David (IBM), Ali (IBM), Mark (SAS)

Discussion:

• Current status of deliverables
  • No status changes to report on RM items since last week
  • Ali pointed out that on the release plan, the dev cycle for this iteration ends on June 4, and the test cycle is 3 weeks. We all agreed that it makes sense to lengthen the dev cycle and shorten the test cycle, given that this is no longer the final iteration for 1.0 and that i11 has a focus on stability. TODO: David to raise this issue on community call. David and Ali to meanwhile ensure that high priority i11 items are completed, regardless of the end of the dev cycle.
• John Arwe, SML spec lead, has provided us feedback on possible items for improvement in testing & implementation, and that should result in some additional bug reports forthcoming
• Mark asked for areas where he could contribute to RM and get his feet wet. David suggested that he could look at bug 229890 as a good starting point for validation work, and Ali agreed. Several of the other defects are more related to Eclipse APIs, and bug 232078 is just grunt work to refactor the package names for our API guidelines. David mentioned that CA might be looking to get more involved with RM as well, given its synergy with CMDBf and their participation in the SML workgroup. TODO: Mark will look at 229890 this week and possibly others, and will work with David & Ali to better understand the workitems and to get pointers regarding the RM codebase.

3 June 2008

Attendees: David (IBM), Ali (IBM), Mark (SAS)

Discussion:

• Current status of deliverables
• Ali said he would only be able to get to RM defects during the test cycle and would submit them as patches, due to his enhancement deliverables. These will all be marked as At Risk.
• Mark is making progress on 229890.
• David's remaining SML defects are At Risk due to other responsibilities as well.

10 June 2008

Attendees: David (IBM), Ali (IBM), Mark (SAS), Merri (SAS), Jason (SAS)

Discussion:

• Current status of deliverables

8 July 2008

Attendees: David (IBM), Ali (IBM), Mark (SAS), Hubert (IBM), Jason (SAS)

Discussion:

• Current status of deliverables

• Introduced Mark to Hubert
• Hubert is setting up SML development environment, and will also read SML specs
• Mark is entering new defects today based on problems and observations he made during i11
• Ali has completed PSVI rework on acyclic builders, validator, and tests. He is modifying document builders to be aware of SML-IF bindings.
• David is working on the substitution builder for the target builders. He also is determining why target instance builder test is failing on IBM JRE. Apparently Xerces needs to be ahead of the JRE on the bootpath. David has the todo to confirm we are legally clear to prereq Xerces.

15 July 2008

Attendees: David (IBM), Ali (IBM), Hubert (IBM)

Discussion:

• Current status of deliverables

Status:

• Hubert looked at removing dependencies for SML MDR. He was able to reduce them, especially in the web services. After that we still have some Eclipse dependencies, most of which are EPL. One of them is an OSGi bundle with 3rd party code. Hubert is checking to see if we need an IPZilla for it. The build script has changed to not package the jars and plugins that are not needed.
• Hubert is looking at locid feature. He has made some progress and is looking for some clarifications, from us and maybe John.
• Ali met with Sandy Gao about the validators. He found that our schema bindings should be interpreted differently, that each schema binding should be treated as one validation set, done one at a time. Any references between the sets should be raised as invalid. This requires changes in our framework code. Ali is trying to minimize disruption to the data builders. The net is we might need to parse through a document multiple times.
• David working on target & schematron validators, reworking to implement and reuse parsing of element type data, and developing new JUnits.
• Overall status of PSVI work: on track
• Hubert presented his thoughts on locid. We determined that there does need to be an implementation of it, that it's not simply coming up with an example. We determined more clarification is needed as to how the locale is specified, and whether substitution needs to be supported. Hubert will follow up with John Arwe to determine this. Status information is pending this decision.

22 July 2008

Attendees: David (IBM), Ali (IBM), Hubert (IBM)

Discussion:

• Current status of deliverables

Status:

• Hubert finished main work for locid. Needs to integrate with HEAD once the schematron code is activated there, and also to write JUnits. Hubert wanted to know what is next in priority. Ali advised moving to the base64 enhancement.
• David will check with Sandy Gao as to the status of the spec (John Arwe is on vacation), as they seem to be wrapping up a new LC version, and whether the Interop will be based on that, and if xml:base support will be part of that. Also to find out status of Microsoft's interop testcases.
• Ali finished most testcases for the work he has been doing. Next he will continue working through TestMainValidatorLC and TestSMLModelUnits. He will be pulled away for other business temporarily for 2 days this week.
• David has finished work on the target validator, except for possible failures to some tests in TestMainValidatorLC (he moved on from that since it was unclear if those tests were valid, due to them being commented out). He is working now on the SchematronValidator.
• PSVI work is still on track for our original estimate.

29 July 2008

Attendees: David (IBM), Ali (IBM), Hubert (IBM), Mark M. (SAS)

Discussion:

• Current status of deliverables

Minutes:

• David working on finishing target and schematron validators. Tests for TestMainValidator pass, but there are failures in the *LC tests. Also need to look at removing an extra builder that might be unnecessary. Will also need to run TPTP tests for i12 iteration.
• Ali temporarily pulled away last week. Now back on RM work, working on existing testcases, expected to be done soon. Then he will move to new testcases.
• David to contact MS (Kumar Pandit) to find out status of testcases for interop.
• Hubert working on merging his locid changes with HEAD. Needs to test with latest code. Also working on base64 enhancement, work in progress.
• Mark hasn't had much time to work on RM, mostly internal stuff. Wants to pitch in, and is open to doing some of the testcases necessary.
• David reminded group of i13 designs needing to be written soon, but it's doubtful there are new designs required for RM.
• Ali to contact Sandy Gao and John Arwe to determine whether our validator needs to support validating schemas that have no corresponding instances. PSVI only operates on instance data, so validating schema-only SML-IF files is difficult.

5 August 2008

Attendees: David (IBM), Ali (IBM), Hubert (IBM), Mark M. (SAS)

Discussion:

• Current status of deliverables

Minutes:

• Mark has updated to get the latest code. Hasn't had a chance to look at the JRE issues yet.
• Ali: most old testcases passing. Creating new testcases for schema binding and other areas.
• Hubert: last week worked on base 64, but having problems with testing. Also did some work merging the locid implementation with HEAD. Hubert encountering problems with running schematron validator. Ali was able to recreate these errors, so David will investigate, since he does not see the errors in his workspace. Hubert also reported that the validator is not correctly handling DTDs, which is needed for base 64 support. Ali suggested using Sandy Gao as a reference as to how DTDs should be handled.
• David: working on finishing up target validator cases.
• Ali: sent email to Sandy asking if schemaComplete must be set to true for validating SML-IF documents that don't contain SML instance documents.

12 August 2008

Attendees: David (IBM), Ali (IBM), Hubert (IBM)

Discussion:

• Current status of deliverables
  • locid and DTD summary
  • test plan status

Minutes:

• Hubert says requirements of locid and DTD (base64) are now clear. He's almost ready to check in his fixes. We need to watch for possible effects of sometimes setting validation to false, and whether we handle schemaComplete="false" (possibly a non-issue). Hubert needs to run testsuite to ensure no regression.
• Sandy at yesterday's meeting suggested we create a DOM representation of the document. David was concerned about the loss of line number information. Ali pointed out that line number information is not a priority for the interop - this is strictly for our problems view in the UI.
• Ali going through the testcases and making sure we have all the tests we need, based on the info received from MS. Also has updated properties files.
• David needs to finish up the target validator, still has some cases that are failing in derivation and same name, after reworking the logic there. Also will change test plan generator to reflect changes made by Ali to metadata.

19 August 2008

Attendees: David (IBM), Hubert (IBM)

Discussion:

• status around the room

Minutes:

• Hubert: bare-names feature - should be an easy thing to complete; also working on problems with sml:ref. yesterday he found a problem with dangling pointers and fixed it. Hubert will be out of the office Wednesday thru next Monday. After bare-names work, he has to do some testcase cleanup, and also needs to change the code to use the latest version of schematron (basically removing some workarounds that were added for the previous version)
• David: since last week's meeting, David has finished the target validator work and has completed some of the testcase changes necessary for the latest spec draft (http://bugs.eclipse.org/244466). Next he will update the export & new SML-IF file create wizard to comply with the latest changes, and will also go through a list of testcase defects that were sent to us by MS. At that point, he will begin work on xml:base (http://bugs.eclipse.org/238492).
• Ali submitted his status via email, as he was unable to attend:
  • Completed:
    • http://bugs.eclipse.org/244307 lax validation when no schemas bound, strict validation otherwise (since related to 244308 incl test cases) [COMPLETE]
    • http://bugs.eclipse.org/244308 add "noSchemaBindings" element [COMPLETE]
    • http://bugs.eclipse.org/244552 change schemaBinding/namespaceBinding from 1..1 to 0..n [COMPLETE]
    • Addressed all incomplete test cases for acyclic and deref
  • Upcoming work:
    • There is currently a problem with the Identity validator he is trying to resolve.
    • Ali will be attending to the incomplete test cases for identity next. There are some cases here where code modification is needed.