2020-01-27 DM SST Agenda and Meeting notes

Leanne Guy

PST meeting held last week  Fri/Sat 24/25 January

• Discussion of the reduction of the commissioning science verification & validation surveys from 5 months to 2 months to increase construction schedule contingency to 4.5 months. The AMCR (previously AMCL) are concerned that schedule contingency might be insufficient but warned against further squeezing of the allocated commissioning period. 
• No guidance yet on the location of the data facility for operations. 
• In-kind contributions: 40 LOI’s have been received, the agencies will assess each of these individually; does not need to be finalized before the Operations proposal goes in.
• Discussion of the impact of reduction in the duration of commissioning science verification & validation surveys on early science with LSST. We might need to modify observing cadence somewhat to mitigate negative effects – still much work to be done on this. 
  • Robert Lupton  thinks we should we should define what we want scientifically rather than just do what we can with can with whatever data is taken in commissioning. 
• Project will deliver a summary report (based on Tony Tyson's work) on the impact of the Starlink satellites to agencies. Expecting ~50,000 satellites by mid-survey (dominated by SpaceX), most in 550km orbits. Most cross-talk can be removed.  Tony needs to give them a number for the factor by which we want the satellites to be darkened. 
  • Robert Lupton  does not agree on the issues of non-linear crosstalk presented. He suggests that DM need to work with Aaron's group to understand the crosstalk signal. He wants the project to sign-off on this report before it goes out.  
    • Colin Slater  is there consensus that these things are not saturating?
    • Robert Lupton  Claim is they saturate at about 10 pixels wide.  
• LATISS is on the mountain; everything is up and running now (took longer than would have liked). Successfully commanding the OCS from Jupyter notebooks, hitting OODS, data getting ingested, etc. Power-up problems discovered on WREBs; took a lot of effort to find the issue; people working hard over the weekend. Hope to be on-sky tonight.
• Proposal to rename LSP to VERA: Vera C Rubin Environment for Research and Analysis. Reviewed with Ops directors (Bob, Phil, Amanda) and Communications. General acceptance of idea.
  • Consensus (and from the DMLT meeting) that this idea should be run by Vera Rubin's family before making any change  
    • Gregory Dubois-Felsmann raised a concern about giving a female personification to a data service component. He asked a few people but none were were worried about this.
    • General consensus that this was something to seriously look into as well how it would be perceived on social media. 
    • Eric Bellm points out that first name has a very difference in connotation. It could be like 'Hey Siri - give me my data'
• Cadence white papers: Over 100 simulated LSST surveys that incorporate optimization suggestions extracted from over 40 white papers submitted by the community, will be delivered in about two to three months. The SAC is forming a Survey Cadence Optimization Committee (SCOC), which will  select and recommend a baseline cadence strategy for the first year of the Rubin Observatory Legacy Survey of Space and Time. The process by which this will happen is still under discussion

• Leanne GuyRobert Lupton talk offline and propose a process by which DM can review and give feedback on  the Starlink analysis prior to release 14 Feb 2020
• Leanne Guy feedback the concerns on the VERA name to the PST and whatever body is charged with making the decision.  (2020-01-28: Concerns raised at subsystem scientists meeting. Zeljko Ivezic will raise at the JOG meeting and clarify the process for renaming)14 Feb 2020 

Eric Bellm

Active branch of DMTN-118: https://dmtn-118.lsst.io/v/DM-19593/index.html

• slides

Comments and feedback around the questions posed on slide 6

• How to choose which features to include: 
  • No resources for a data challenge. Plasstic , which was much smaller and more contained took a lot of effort. 
  • Discussions with users and at meetings have not so far resulted in a concrete list of what is needed. General consensus that the DM-SST should make a best effort list of what will cover all science needs, present to the science collaborations and use the successive pre-operations Data Previews (DPs) to assess whether they meet user needs 
• Should AP & DRP have the same features?
  • General consensus to  accept the possibility that they can be different. 
  • Leanne Guy suggests they should be drive by the science goals of AP and DRP respectively. 
• Should we compute features on difference or total flux (or both)?
  • Yusra AlSayyad states that DRP was planning to put effort into this with HSC data to understand which option we should uses, are they consistent and what at the characteristics of each are. 

• Will LSST compute the same features over the ten year survey?

  • It is a science requirement  that we need to be able to change things. 
  • DRP can easily change the way the DB looks with each DR. The PPDB is less flexible. When/how can we update the PPDB schema that would not  be invasive. 
  • Important to make clear the expectation that

• Gregory Dubois-Felsmann notes that so far only the scientific aspects of the features have been discussed. We also need to discuss the query interfaces,  IVOA interfaces and Python wrappers for these interfaces. Maybe this will be a separate note but do need to try to make a  concrete statement about of this.   

Work in progress - Eric now actively soliciting feedback from the DM-SST. Will accept branches in github, unstructured text via email. 

• Gregory Dubois-Felsmann and Eric Bellm to discuss what can/should be said at this point about IVOA and Python interfaces to querying time series data in both the Prompt and DR data products. 14 Feb 2020 (deadline set to before the DMLT virtual-F2F to enable followup there, if needed)

Leanne Guy

• Keith Bechtol: outlines for LSST construction papers due this week. Number of papers on related topics.
• Leanne Guy : Zeljko Ivezic wants an outline of each first and to use that as a basis for possibly merging papers.