Date: 9 August 2010

Agenda:

• spec issues resolved since last workgroup meeting
• Collect any additional feedback from work group on current draft spec
• Core spec compliance table added to Compliance section.
• Clarifications in Resource Formats section about RDF and XML content types

Minutes:

• Reviewed spec issues
  • The workgroup would like to be more aggressive in deprecating the QM V1 specification. Currently, the QM V2 spec requires the client to send the OSLC-Core-Version header. If the header is not sent then the service provide would assume that the client wishes to use QM V1. A more aggressive stance would be to require QM V1 clients to explicitly request the media type defined in the QM V1 spec (e.g. Accept: oslc-qm-*-1.0). This more aggressive stance might force clients of the V1 spec to patch, but would have the advantage of allowing clients to get QM V2 content without providing the OSLC-Core-Version. At this time there are two known consumers of the OSLC V1 spec mentioned in the Implementation Reports: Rational Team Concert and Rational Requirements Composer. If these consumers can adopt the change then the QM V2 spec can reflect the more aggressive stance.
  • The "occurs" column is improved since last meeting but the TestScript/executionInstructions property now has "unspecified" value in this column. Workgroup discussed whether the value should be concrete instead of unspecified. It was decided that some service providers may require this property in order for the TestScript to be funcational while other service providers may not be able to provide a value for it. Therefore the "unspecified" value will be left in place and a comment will be added to the spec to clarify that cardinality is provided in the resource shapes document.
  • AI: PaulMcMahan to create an issue
• The Compliance section does not match some of the details later in the spec in terms of MUST/SHOULD. If a row is marked as MUST then the corresponding language later in the spec needs to be double-checked to be sure that it aligns.
  • AI: PaulMcMahan to create an issue
• Workgroup discussed whether to enter a final convergence stage where the only changes to the spec would be for defects. It was agreed that the time seems right to enter final convergence but it should be discussed via the qm mailing list to give workgroup members that were absent from the meeting a chance to weigh in.
  • AI: PaulMcMahan to send email on final convergence to the QM mailing list