Project Meeting 2020.10.01

Task 1 Project Management (milestone)

• For the technical call, we'll set an agenda ahead of time and folks are encouraged to add to it. We'll also note who attends.
• Discussed 5b and 6a amendments
  • @Bill email no cost extension of 5b to end of year
  • @Bill email NTP for 6a with an end date of June 30, 2021
  • Jeff will finish work on 5b before getting into 6a tasks
  • I'll start on 6a - getting the SEMCOG related tasks kicked off and getting Jeff Newman under contract
  • Let's do a better job of memorializing revisions to scope in 6a and beyond
• Lisa helping with PopulationSim non-binary incidence feature review
• There's a dependency update issue that we should discuss at the next technical call #349 and #350
• Jeff fixed the joint tour chunking issue #288 discussed Tuesday
• Need to review Clint's ARC parking location choice work and discuss Tuesday

Task 2 Strategic Planning (milestone)

• Discussed comments on the multiple models section of the updated version of the plan
• Earlier draft version with comments here
• We're going with the few core model designs / pipelines / spines approach to example/test management as opposed to adding every region
• With that being said, I'll add a paragraph about an optional plan for testing agency contributed models as well
  • It would be optional and paid for separately
  • It would be coordinated / run by the team
  • It may be run nightly/weekly instead of with each commit
  • It would require private (paid for) data management and computing solutions on top of the existing free open source tools being used
• I'll add a development roadmap to the contributions from others section since we want to be more explicit about the project's future
• @PMC can review the updated strategic plan next week

Task 10 Support for Three Zone Systems and TVPB (milestone)

• Jeff finished adaptive chunking, which will speed up existing models and help with TVPB memory management
• Jeff now setting up the marin tm2 work tour mode choice example
• After that, work on caching / parallelization
• Should we do some more performance tuning (such as better manage data types) since Jeff is into that part of the code now? Since TVPB leads to some memory issues, should we maybe revise our approach? Let's discuss on Tuesday.