wiki:JeffsNotes
Last modified 10 years ago Last modified on 07/16/2009 05:56:13 PM

SDQA Management Session June 11 (Jeff, Tim, Gregory, Roc, Deborah, Robert)

Jeff: Tim owns responsibility for looking at data coming out of data challenges. Tim put together plan how we will do DC3b science data validation.

Jeff: IPAC needs to develop system design that will make the process work. Tim = requirements, Deborah = design/implementation.

Deborah: Tim defines what he needs for DC3b, and IPAC continues to address volume and ops concerns. Why are these different? (Tim)

Jeff: DC3b scope for SDQA is to validate data outputs. It can also be used to validate pipelines.

Tim: Examine processing of individual images as well as assembling into overall data release and examining that.

Jeff: Need continuous early integration.

Roc: Pure pipeline validation vs data release meeting DC3b requirements (ie SDQA framework can be used for both).

Jeff: Robert to have developers come up with items/metrics needed to state that pipeline is processing correctly.


Jeff on white board-

  • Tim must produce DC3b requirements, DC3b validation plan, and must do validation. Tim to engage Deborah in this work with an eye toward operations, scale. Assisted by other scientists. This will drive user interface tools.
  • Robert will ensure initially and repeatedly that all pipeline stages have validation driven metrics.
  • Robert will ensure pipeline stages get integrated into harness early and often.
  • Deborah translate these (both data validation and pipeline validation) into SDQA system features. Document next level of detail in UML model.
  • Tim is prime initial customer for SDQA system.
  • Applications manager responsible for determining if scope fits in DC3b.