Keith Bechtol has a grad student access - what level of access can be provided to LSST data and services? Keith is paid by LSST (so is staff and has full staff privileges), the grad student is not. There are (currently) 2 granularities of permission, LSP (lspdev) access and full staff. The understanding seems to be that full staff privileges can be given if Wil O'Mullane see:  DMTN-020

• Simon Krughoff to check with Wil the policy for giving data access to grad students 31 Aug 2018 

Leanne Guy Robert Lupton

DESC are starting DC2 soon. Several meetings have been held between DESC and DM Science to see how we can work together:

• Joint hire to work on algorithms, preferably based @ princeton
• Hosting of DC2 data at NCSA: Main motivation from DM is supporting QServ development.
• DM processing of DC2 Run 1.2i: Testing of LSST-like data and support for QServ testing
• Image differencing pipeline: DESC has 2 people optimistic about contributing to the source association part

Run 1.2i is ~ 1/3 the size of the HSC dataset. It was noted that Jeff Kantor will be looking for a similar sized dataset to use for network transfer tests between La Serena and NCSA and might use this dataset.

Eric Bellm has discussed the DIA pipeline source association with DESC - has not concluded. Does not understand why what DM has is not enough. Eric and John will discuss further.

Robert Lupton says DESC has discovered that we don’t have tools to match associating sources to objects and sources, e.g across bands - 25 images on sky in 3 bands. This triggered a discussion on what representations are we using for data? FITS, parquet? Simon will follow up

Jim Bosch raised LIT-94 - Getting issue details... STATUS again w.r.t engaging DESC to help with validation of shear estimation requirements. Does the LCR depend on successfully enlisting DESC to help with validation? RHL is concerned about including shear estimation requirements if we are not scoped to do the validation. How can we maximize the interaction with DESC on this subject. Will discuss in more detail in 31 Aug 2018 SST meeting.

• Leanne Guy to follow up with Jeff Kantor about possibly using the DESC dataset to be hosted at NCSA for network tests 31 Aug 2018 
• Leanne Guy to come up with a plan for LIT-94 - Getting issue details... STATUS with Jim Bosch 14 Sep 2018 
• Eric Bellm to discuss/review DIA pipeline with John Swinbank 07 Sep 2018 
• Simon Krughoff to follow up on the plans for data representation 21 Sep 2018 

Leanne Guy

There are several open tickets on the DPDD, some need further discussion and clarification. All changes to the DPDD must go to LCR. Plan is to include all (if possible) in the upcoming LCR.

Issues concerning PPDB tables were raised in DM-13375. Will ForcedSources be added to the DIASource table with a flag,. if we break out a DIAForcedSource table we should define what we will measure. Gregory asks would a DIASource record be created for retroactive DIAForcedSources - could be confusing? Could create different views on the one table. Eric will RFC these points separately. DM-13375 will be completed as originally defined.

LDM-153 is referenced from the DPDD as the source of the DPDD schema. It is currently not consistent with DPDD. GPDF noted that the schema should be secondary to what we want scientifically. Who maintains LDM-153 now that Jacek is gone? The lsst/cat package contains a machine readable description of the schema. The text in LDM-153 looks to be copied from lsst/cat. There is a preference for a machine readable package as opposed to a document. Can LDM-153 be autogenerated from lsst/cat? Ask Colin Slater to follow-up

RHL expressed unhappiness with the Source table. This is a larger discussion. Robert Lupton will chair a session on this subject at the SST-F2F meeting in November at Princeton.

• Colin Slater To discuss with KT and Margaret (in the context of updates to the sizing model) their preferred form for the DPDD schema (document, lsst/cat package?) 07 Sep 2018 
  
  
• Eric Bellm to create an RFC on the topic of PPDB tables, DIASource/DIAForcedSource 31 Aug 2018 
  

Robert Gruendl Simon Krughoff

Robert Gruendl reported on progress on transferring camera test data to NCSA. Good progress is being made. There is an ongoing discussion about what needed to be (should be) copied. General consensus is that is is just easier to move everything and sort it out later - disk is cheap. Previous issues concerning mutable data are understood - the data themselves are not being modified, just the definition of which data comprise a given test.

What are the plans for eTraveller? Can information in eTraveller be moved to the EFD? Do we replicate eTraveller at NCSA or is it best to point a python interface to eTraveller?

AuxTel testing has started in Tucson and they are facing the same issues. Robert L and Leanne are working with Patrick and NCSA to make the spectrograph data available at NCSA via the LSP with an acceptable latency for testing.

• Leanne Guy Robert Lupton to work on automatically moving spectrograph data in Tucson to NCSA and making available via LSP 07 Sep 2018 

Leanne Guy