Hi everyone,
Apache Beam entered incubation in early February. Over the past 10 months,
the podling has made great progress across various areas: refactoring the
project to remove any special treatment given to a runner or a vendor,
building processes that encourage open development, evangelizing the
project, and growing the community.
Now, with the support of our mentors and overwhelming support from the
wider Beam community [1], I’d like start a discussion on the progress we
have made and a possible graduation recommendation as a new top-level
project.
To prepare for the discussion, we have published our self-assessment [2]
against the Apache Maturity Model. We tried to include links and evidence
whenever applicable. I’ll summarize the commonly asked questions here, but
please see the self-assessment for additional information, various details,
graphs, evidence, etc.
> Releases?
Three -- all unanimously approved, all driven by different release
managers, across different organizations. A detailed release guide is
available on the website.
> Community growth?
There has been a clear growth month-over-month. We have had 1500+ pull
requests on GitHub and 110+ individual code contributors. In terms of
mailing list activity, over the past 30 days, we have had 50+ individual
participants on dev@ and 35+ on user@.
> Organizational influence?
We have worked hard to remove any special treatment given to any
organization. The bulk of the initial code donation came from Google, but
now both the project’s code and branding have a clean separation between
the project and Google Cloud Dataflow (which has become just one of many
runners that can be used within Beam).
While it is true that Googlers continue to provide the majority of commits,
over the last three months no single organization has had more than ~50% of
unique monthly contributors. (Please see the graph in the self-assessment.)
Diverse influences are also particularly clear when you look across modules
within the project. Beam has about ~22 large modules in the codebase, at
least 10 modules have been developed with little to no contribution from
Googlers.
Now, if we were to graduate, the Beam PPMC recommends the following
information for the Board resolution:
* Project name: Apache Beam
* Project description and scope: a unified programming model for both
batch and streaming data processing, enabling efficient execution across
diverse distributed execution engines and providing extensibility points
for connecting to different technologies and user communities.
* PMC composition:
* Tyler Akidau <taki...@apache.org>
* Davor Bonaci <da...@apache.org>
* Robert Bradshaw <rober...@apache.org>
* Ben Chambers <bchamb...@apache.org>
* Luke Cwik <lc...@apache.org>
* Stephan Ewen <se...@apache.org>
* Dan Halperin <dhalp...@apache.org>
* Kenneth Knowles <k...@apache.org>
* Aljoscha Krettek <aljos...@apache.org>
* Maximilian Michels <m...@apache.org>
* Jean-Baptiste Onofré <jbono...@apache.org>
* Frances Perry <fran...@apache.org>
* Amit Sela <amits...@apache.org>
* Josh Wills <jwi...@apache.org>
(The ratification of the full text of the draft resolution is nearing
completion.)
While we have made great progress across the board (thanks to so many of
you in this community), I’m sure there’s still plenty to do. We continue to
be focused on community growth, and processes that encourage open
development. Regardless of the outcome of this discussion, we’d love to get
specific feedback on what can be improved going forward.
Any thoughts, comments, questions or concerns? Thank you.
Davor
[1]
https://lists.apache.org/thread.html/f133fb6bf2d1851d1bd5880c772e4b050700154fa178fdb00a5b66bf@%3Cdev.beam.apache.org%3E
[2] http://beam.incubator.apache.org/contribute/maturity-model/
