- what is pluggable table storage
- architecture
  - Diagram of an update statement (parser -> planner -> executor -> indexam SELECT -> tableam SELECT -> executor -> tableam UPDATE -> executor)
- what can you do in v12
- who worked on it
  - Andres Freund
  - Haribabu Kommi
  - Alvaro Herrera
  - Ashutosh Bapat
  - Alexander Korotkov
  - Amit Khandekar
  - Dmitry Dolgov
  - many more

- which areas are lacking
  - tuple identifiers
    - index restructuring
    - don't constantly copy tids / pass-by-ref
    - useful for:
      - global indexes
      - AMs with wider tuple identifiers, including IOT
    - hardest part: indexes
    - other challenge: core planner / executor
  - planner / executor integration
    - in particularly partial accesses at scan time, should know which
      columns are going to be needed (especially important for column
      stores, but also relevant for row stores)
  - function naming and other structural issues
  - DDL extensibility
  - analyze integration