This was a week of updates from the OpSim working group.
Cathy and Kem reported on the reference run document update. The document is not done yet, but should be done soon. The contents of the CCB document and other inputs are in Docushare Collection 4604 - https://docushare.lsstcorp.org/docushare/dsweb/View/Collection-4604
Cathy ran through the changes in terminology - 'baseline cadence' is overloaded, but should be recognized to refer to the overall set of proposals + expected observing strategy in the reference run. Perhaps "baseline observing strategy" is more clear. "Universal cadence" is discouraged, in favor of "WFD cadence", which refers to multiple (not necessarily just 2, although that is what is in the current reference) visits inside one night. The footprint for the WFD cadence can vary .. in practice, in the baseline observing strategy, has WFD cadence in the WFD footprint and also in NES footprint, for example. Cathy will write up a community post clarifying this for the community.
Opsim3_61 to minion_1016:
- skybrightness is the same (will change in the future)
- scheduled downtime has increased
- m5 expected values changed due to seeing + Cm values changing (not going as faint)
- went from stretch/stretch to design/design footprint and number of visits for survey
- more slewing if we want bigger moon avoidance angle
- generally .. more realistic and less optimistic
Michael updated us on SOCS/scheduler status.
- v0.3 released in May - first real implementation of (a single) proposal, that allows scheduler to serve real targets through the system. allowed running of observatory models in SOCS and scheduler (scheduler has ideal system, SOCS has "real" system that receives observation, does slew, takes image, etc.) - gives same result for at the moment, although plan is to allow SOCS to add variations against ideal (allow degradation of telescope over time).
- Now cleaning up sky regions related to proposals - adding more proposals.
- Francisco adding time-dependent proposals.
- Michael will do first pass at downtime before SPIE.
- Francisco and Michael on vacation until mid July, then likely to be busy with Project & Community workshop.
- Need to look at how to validate ahead of next release - work with MAF team to make sure that MAF can analyze the outputs.
- v1.0 release: scheduled for late October. Area distribution (automatically? uniformly distributed in time), time-dependent proposals, new weather - clouds and sky, new methods to rank targets.
- summary table will be in v1.0, but will be a view not a join. formats of outputs are still under change.
- scheduler components can be pulled apart and run separately.
- output of v1.0 may be mysql or sqlite : schema are slightly different between the two, not sure what it will be yet.