Thank you for the responses! I absolutely love the idea for a release party. Let's do it after step 12 - IPMC Vote? Henrik, Can you please organize it?
Meanwhile, we are at currently at Step 4 - Verify supported Python versions. I am going to finish this step this weekend (and open a discussion to switch from poetry and to uv, which makes this and other things easier). Best, Alex On Wed, Apr 2, 2025 at 4:18 PM Austin Bennett <aus...@apache.org> wrote: > Glad this is coming together! Eager for otava versions to get sorted so > can adopt. > > On Wed, Apr 2, 2025, 3:48 PM Henrik Ingo <hen...@nyrkio.com> wrote: > > > Thanks Alex! > > > > I was operating under the assumption that everyone else here knows the > > Apache process better than me, so I was hoping someone to start exactly > > this email! > > > > I have two additions: > > > > - I suggest we do a zoom or teams release party? In fact, we could > > literally do the svn commands together on such a call? I would invite > > people who contributed over the years, all the way back to William Brown, > > who wrote the first version of the algorithm during his last week as a > > Summer intern at MongoDB. (And we never heard from him since but I > recently > > tracked him down on Linkedin.) > > > > - I have written a paper on the evolution of performance improvements in > > Otava from the beginning. I want to publish this partly as a homage to > > everyone who improved it over the years. Spoiler:Original implementation > > was literally what the e-divisive paper describes: O(n^3) and the next > week > > William himself improved to O(n^2). Last year I achieved constant time > > (when adding a single point). In raw numbers the performance improvement > is > > about 300 thousand times faster :-) (Depends on amount of data and also > > number of change points found.) Main performance improvements came from > 4 > > different developers over 8 years. > > > > - I might and probably should accompany the above with a more casual blog > > post telling the history of Otava from MongoDB to Datastax to ASF. (And > > other notable users too, of course.) > > > > - I will be presenting at > > > > > https://icpe2025.spec.org/tracks-and-submissions/research-presentation-track/ > > on May 7-9. ICPE is the conference where both MongoDB and Datastax > > primarily published papers on Otava / Hunter / Change Point Detection. > > > > > > henrik > > > > On Mon, Mar 31, 2025 at 1:35 AM Alexander Sorokoumov < > > aleksandr.sorokou...@gmail.com> wrote: > > > > > Hey folks, > > > > > > I think we are very close to cutting the first Apache release. Yet, > there > > > are many small steps to do. Personally, I have been procrastinating on > > them > > > as I don't exactly know what to do. > > > > > > To get over this "fear of the unknown", I have made a painfully > detailed > > > TODO list that should get us to the first release. The purpose of this > > > thread is to a) agree that the steps in this list will indeed get us to > > > correctly publishing the first Apache release and b) clarify the > > unknowns - > > > mostly interactions with Apache infrastructure. > > > > > > I do believe that once we agree on the outstanding steps, we'll be able > > to > > > get it over with within a week or so. > > > > > > Please find the list below. My unknowns are marked with **Q** and > ????. I > > > am looking forward to reading your feedback! > > > > > > TODO towards the first Apache release. > > > > > > 1. Agree on the release version (0.6.0-incubating? I am using that as a > > > placeholder in other steps). > > > 2. Update release name in pyproject.toml. > > > 3. Update project description and authors in pyproject.toml. > > > 4. Verify supported Python versions. Run tests and examples with Python > > > 3.8+. Add a paragraph to README claiming that versions where tests > passed > > > are supported. > > > 5. Create source distribution via `poetry build`. > > > 6. Sign the artifact `gpg --armor --output dist/<artifact>.asc > > --detach-sig > > > dist/<artifact>`. > > > 7. Verify the checksum `sha512sum dist/<artifact> > > > > dist/<artifact>.sha512`. > > > 8. Test release candidate locally with `pip install`. > > > 9. Upload project KEYS file. **Q:** How do we do that? > > > 9. Upload Artifacts to Apache Dev Repo. **Q:** How do we do that? > > > 10. Call for PMC Vote. Send an email to the this mailing list with the > > > subject "VOTE: Release Apache Otava 0.6.0-incubating" with the link to > > the > > > release candidate: > > > ``` > > > The Apache Otava project is pleased to announce that the > > > release candidate for version 0.6.0-incubating is ready. > > > > > > The release candidate artifacts are available at: > > > ???? > > > > > > SHA512: <checksum> > > > KEYS:https://downloads.apache.org/incubator/otava/KEYS > > > > > > Please vote on releasing this candidate by replying with: > > > [ ] +1 Release this package > > > [ ] 0 No opinion > > > [ ] -1 Do not release (please provide reason) > > > > > > This vote will be open for at least 72 hours. > > > ``` > > > 11. Wait for the vote to pass, fix issues if any. > > > 12. Call for IPMC Vote. Send an email to gene...@incubator.apache.org > > with > > > the subject "[VOTE] [RESULT] Release Apache Otava 0.6.0-incubating". > The > > > content should contain links to the vote result and the artifacts. > > > 13. Wait for the vote to pass, fix issues if any. > > > 14. Move the RC to the release directory. **Q:** How do we do that? > > > 15. Publish Github release/tag. > > > 16. Upload the release to PyPI. **Q:** How do we do that? > > > 17. Announce the release on the mailing list by sending an email with > the > > > subject "[ANNOUNCE] Apache Otava 0.6.0-incubating released". > > > 18. Make social media noise. > > > 19. Bump version in pyproject.toml for the next release. > > > > > > > > > Best, > > > Alex > > > > > >
