SAT Meeting 2014-07-08

• (on travel)
  

• (on vacation)
  

• (on travel)
  

• (at SciPy 2014)
  

• Discussed Level 1: Level 1 Calibrated Exposure Processing and Level 1 Difference Imaging and Moving Object Processing
• Is ISR suitable for wavefront sensors? (They're out of focus, among other things.) Gregory to ask Chuck (Bo/Srini).
• Unclear how well we need to know the PSF for L1.
• We can use data from L1 or a previous DR to set initial conditions, but provenance points to that specific data.
• Use multifit outputs for DCR-corrected template generation?
• Filtering false positives?
• Where/when do glints and other artifacts get masked?
• Where do fakes get inserted and how?  And how do we make sure they're flagged appropriately in the outputs?
  • Do we have to insert into the template as well as the image?
  • Do we have to process faked regions twice?

Data Storage

• Conditions/calibrations (bitemporal data):
  
  • Butler handles all queries.
    • But we need technologies underlying the butler.
  • Gregory has a preference against numbers as code, but doesn't take git into account.
    • git is not a conditions database.
  • For data intended for tests, there is a third dimension beyond bitemporality: versioning.
  • Need to have at least a metadata database, even if data itself is not in database.