• Approved by acclamation.

HDB Well-Formedness.md document is live under "technical-documentation".

Aaron to go over what it is and what it isn't.  Hopefully opens the door to shared development of HDB utils.

• Aaron unable to make the call.
• Good discussions about extensions that go beyond Kx functionality.
  • Ideas that have been implemented at Morgan Stanley:
    • Sparse HDB
      • If new table is introduced, backfilling many years of empty table segments is time consuming and makes schema changes more expensive.
    • Staging Directory
      • When a table segment is only partially written due to I/O speed limitations, the HDB as a whole is technically not well-formed.
      • Staging a table on the same mount point and moving it atomically when writes complete could avoid problems with intra-day HDB reloads.
    • Partial HDB Reload
      • Intraday writedown (w.q) often implies intraday reload.  Loading of HDBs is I/O intensive.  A means of loading for a target date range could help tremendously.
  • JP Morgan interested in extensions that support archiving in a standard way.
    • Perhaps recent data run on more performant storage, whereas older data would be hosted on slower, cheaper drives.
• Perhaps worth creating an rfc/ directory in the repository to get the details out in a shareable format.

• Should be able to turn reviews / approvals around more quickly.
• Hope to add one or two more based on strong contributors.

Managed to merge a couple of branches to main (with the help of Magit):

• structlog

• authz_ro

Hoping it makes it easier for others to improve on them.

• timer pull request to be merged soon.

• memusage pull request is up.
• Question arose about how documentation should be maintained.
  • AquaQ is familiar with GitHub microsites as they use those already.
  • Suggestion that they talk to Eric Griswald to see what (if anything) is different with microsites under FINOS compared to what AquaQ is doing already.