Date: Thursday, 27 October 2011

Time: 1:00 PM Eastern US (contact MichaelFiedler if you'd like to participate)

The Automation meetings alternate times each meeting to accomodate the global team.

Agenda

* Reoccurring agenda items:

• Recap of previous meeting AutomationMeetings20111020

• Continue review of the automation execution scenario
  • Overall flow - applicability to workgroup member scenarios
  • Issues with passing required parameter/environment information during Automation Job creation
• If time allows, discuss contributions to automation results
• Plans for moving from scenario development to spec development
• Previous Action Items:
  • NA
• Next meetings:
  • 3 November

Minutes

Attending: Michael Fiedler, Dave Brauneis Bill Higgins, Daniel Berg, Paul McMahan? , Vaibhav Srivastava, Charles Rankin, Rich Rakich
Let MichaelFiedler know if anyone's name is missing.

• Discussion focused on automation execution scenario
  • Need some additional context to evaluate scenario completeness. Put it in context of build, deploy, test, provision.
    • TODO: PaulMcMahan, VaibhavSrivastava and PramodChandoria will document a test scenario for the next meeting
    • Simple scenario showing a single test execution flow - what inputs needed, what results would look like in terms of contributions
    • Possibly a secondary scenario showing automation chaining. Example: test tool consumes a build automation and starts a test execution automation
  • Re-visited the discussion on whether and intermediate Automation Request artifact is needed.
    • Artifact would link to the Automation Plan it was created for, link to the Automation Results for the job
    • Some members felt a single Automation Job could not capture all aspects of requesting an automation and recording the results.
    • Automation consumers may not have all of the knowledge required to fully create a job. Example: Build requester could have no knowledge of build contributions such as JUnit tests
    • TODO: DavidBrauneis, CharlesRankin and MichaelFiedler will take this offline to investigate pros/cons of both approaches and report back to the workgroup.
  • Discussed the environment/parameter issue documented in the scenario
    • Strong sentiment to keep it simple initially.
    • Alternative discussed: similar to what AM does, expose resource shapes generated dynamically based on registered artifact types. Complex.
    • Alternative discussed: inputs are a simple combination of dcterms-like attributes and possibly an opaque attribute for arbitrary input data.
  • Discussed security aspects of the scenario
    • General agreement it should be explicitly out of scope for the specification, but issues around identity and authentication w.r.t automation should be acknowledged.
• Discussed topics for next workgroup meeting
  • Results of test automation scenario elicitation
  • Agent/Worker sub-scenario
  • Beginning to document artifacts and early specification drafts.

Next meeting: Thursday, 3 November at 1PM Eastern US time