6th October, 2021

Attending: Josh, Ryan W,. Greg Lee, John K., Ward F., Hailey J.

  • Planning Poker, Life, Etc.

  • Ward: software sustainability for NSF solicitation (3yr)

    • Pitching: adding parallelism to C (a la MPI) and adoption of

      V3 (plus educational materials, etc.)

    • Looking for collaborators. (Allen Cell Science, Janelia, …)
  • Josh: moving forward on a couple of fronts

    • Discussion with Norman around prototype for sharding

    • Discussion with xarray contractors around formalizing


      • multiscale, _ARRAY_DIMENSION, tables (recent hackathon)
    • Ward: have xarray compatibility on the main branch to be

      released soonish (maybe in 4.8.1).

      • Ryan Abernathey has recently joined the steering committee.
      • Good to be part of those discussions
      • Introducing new functionality – bit grooming (lossy compression) – improving compression rates by chopping of bits smartly
      • cf. having the burden of enforcing e.g. cfconventions at the right location
      • What is a convention & what is a technical obligation for people working with/on zarr.
    • Josh: best idea so far is new repo that publishes to github


      • Ward: publishing best practices/guidance

        • “Not rigidly enforced by software”
      • John: balance between extension & convention

      • Ward: extensions? see: /core-protocol-v3.0-dev/protocol/extensions.html

      • cfconventions are lexikon of metadata & a process for using them

      • e.g. precise terms to use for labelling data.

      • Josh: in comparison, extensions can change the behavior of code (`timestamp`s returned rather than `str`s)

      • Ward: map into/out of HDF5 for storing netcdf4 such that the HDF5 may not do you any good (without the netcdf library). This feels somewhat similar.

  • Greg:

  • Hailey

    • PR within the next week for codecs in netcdf library.

    • Ward: responsibility for netcdf-java in THREDDS…

  • ZSC: sign the document!

  • Outreachy update

    • accepted, 4 projects proposed

    • no contributors have showed up yet

    • might write up some of the even simpler things (survey)