Le 05/01/2021 à 21:57, Wes McKinney a écrit : > It's time for our quarterly board report (due next Wednesday). Can > everyone help with completing the various sections? > > ## Description: > The mission of Apache Arrow is the creation and maintenance of software > related > to columnar in-memory processing and data interchange > > ## Issues: > [Insert your own data here]
- After our past issues around scarce CI resources, we had managed to stabilize our CI setup around the relatively generous execution facilities provided by Github Actions, and a bunch of additional builds on other platforms (mostly for nightly builds). That equilibrium was unfortunately shattered after INFRA suddenly disallowed most third-party Github Actions helpers in response to security concerns. Our remaining Github Actions builds also seem to be queued longer. Concurrently, Travis-CI also started to be much stricter with resource consumption, which is currently affecting several of our nightly builds until we manage to migrate them to other CI services. These factors compounded mean we'll finally need to provide and set up our own CI resources. We hope to be able to migrate a significant number of builds to Buildkite, using execution resources paid by our sponsors. We also hope that INFRA rules around repository access and configuration won't stifle that endeavour. - The weight of our release procedure (due to the number of implementations and platforms involved) puts a burden on our capacity to release often. Our goal is to cut a new feature release roughly once every two months, which often ends up challenging. This situation also makes us reluctant to put out bugfix releases in-between, except for critical fixes. > ## Membership Data: > Apache Arrow was founded 2016-01-19 (5 years ago) > There are currently 54 committers and 30 PMC members in this project. > The Committer-to-PMC ratio is 9:5. > > Community changes, past quarter: > - No new PMC members. Last addition was Francois Saint-Jacques on 2020-03-04. > - Andrew Lamb was added as committer on 2020-11-06 > > ## Project Activity: > [Insert your own data here] - We'd like to release Arrow 3.0 around mid-January. The issues discussed above (diminished CI coverage due to disallowed third-party Github Actions, blocked nightly builds on Travis-CI, complex release procedure) make it more difficult than we'd like. - There seems to be growing interest in zero-copy zero-dependency cross-runtime interoperability using our "C data interface". It is now concretely possible to share data in-process between R, Python and Rust, without any shared code and without going through IPC mechanisms (the latter still possible, of course, using the Arrow IPC format). Other projects such as DuckDB have started using this interface. - The Arrow in-memory format is regularly being adopted and/or adapted to in other projects, especially in the dataframe and database communities. > ## Community Health: > [Insert your own data here] The project is overall very active. The Rust implementation is especially thriving, with numerous pull requests being submitted and merged in due time. The C++ implementation is advancing at similar speed as in previous reports. A Julia implementation was recently contributed. Our pull request backlog has also grown slightly as a result of increased activity, but it does not seem to be posing a real problem for now. Regards Antoine.
