Thanks Constance for all the summaries and different tallies (yes, it's easy to lose track in mailing list - been there done that).
One of the reasons ASF is working on Apache Trusted Releases - where release votes will be collected by the "tooling" rather than relying on manual processing of mailing list output (ATR will generate the mails so that we can keep track of the votes). BTW. Daniel, also I would really encourage you - if you think it's worth it - to re-run the vote using IRV (instant runoff-voting). There is a good reason you raised this vote could make people confused or unclear on how they should vote. If you continue worrying about that - one way of dealing with it is to re-run the vote, with the reasoning that indeed - people could be confused. I think we all hear and understand concerns you have. Taking matters in your hands to "improve things" and do it "better" is something I would very welcome - and I guess a lot of us would as well. I also see it as a very interesting experiment - It would be great to see what will be the result of it and compare. J, On Wed, Oct 29, 2025 at 4:38 PM Constance Martineau via dev < [email protected]> wrote: > Jens, I am so sorry! I double checked and am not sure how I missed that. > Thank you for catching my mistake! > > Doesn't change the outcome but official updated tally: > > > - Option A: +1.5 > - Option B: +8.4 > - Option C: +5 > - Option D: +3.5 > > > > On Tue, Oct 28, 2025 at 5:17 PM Jens Scheffler <[email protected]> > wrote: > > > Thanks Constance for driving this. > > > > Looking at the Google Sheet it seems my (binding) vote is missing but no > > big problem as I also voted for Option B > > > > On 28.10.25 18:32, Constance Martineau via dev wrote: > > > Thanks Daniel. Not sure how to make them show up, so made the google > > sheets > > > public: > > > > > > https://docs.google.com/spreadsheets/d/1sNhlNM2YqgTDvWOXp7o0zFF-VquddWANxx7S02G3lXM/edit?gid=0#gid=0 > > > > > > > > > > > > On Tue, Oct 28, 2025 at 1:18 PM Daniel Standish via dev < > > > [email protected]> wrote: > > > > > >> Images did not work > > >> > > >> On Tue, Oct 28, 2025 at 10:12 AM Constance Martineau via dev < > > >> [email protected]> wrote: > > >> > > >>> Hi all, > > >>> > > >>> Thank you for your patience while I tallied the votes! For reference > > >>> purposes, here > > >>> <https://lists.apache.org/thread/7mbztc6dchh73c7cnn7sjm1qtt6gj5zw> > is > > a > > >>> link to the vote thread. > > >>> > > >>> As a reminder the options were: > > >>> > > >>> > > >>> - Option A: Prefer dag in docs; use DAG only when referring to > the > > >>> class/import > > >>> - Option B: Prefer Dag in docs; use DAG only for the class/import > > >>> - Option C: Keep DAG as the standard everywhere (status quo) > > >>> - Option D: Prefer Dag in docs, use Dag for class/import and > alias > > DAG > > >>> (for backcompat reasons) > > >>> > > >>> > > >>> *Results (Binding Votes Only)* > > >>> > > >>> Based on the results, and following the rules that only binding votes > > are > > >>> counted, with voters able to submit a fractional vote between -1 and > +1 > > >> per > > >>> option, Option B (Dag) won. > > >>> > > >>> We will therefore prefer "Dag" in docs, and use DAG only when > referring > > >> to > > >>> the class or import itself. > > >>> > > >>> [image: image.png] > > >>> > > >>> *Additional Context* > > >>> > > >>> Because this vote was about convention (not code or architecture), > and > > >>> because the discussion around voting method itself was interesting, I > > >> ran a > > >>> few "what-if" tallies to see how the outcome might vary: > > >>> > > >>> - If all votes (binding + non-binding) were counted, including > > >> multiple > > >>> options per person, Option B (Dag) still wins, but by a much > closer > > >>> margin. > > >>> > > >>> [image: image.png] > > >>> > > >>> - If only the main binding vote (single strongest +1 per voter) > > were > > >>> considered, Option B (Dag) and Option C (DAG) would have been > tied. > > >>> > > >>> [image: image.png] > > >>> > > >>> - If the main vote from both binding and non-binding voters were > > >>> included, Option C (DAG) would have narrowly won) > > >>> > > >>> [image: image.png] > > >>> > > >>> (For transparency, Ryan Hatter submitted two +1s, so I pinged him to > > >>> clarify why should be considered in the single-vote scenario) > > >>> > > >>> *Observation* > > >>> It's interesting that the outcome differs slightly between binding > and > > >>> non-binding voters, with contributors leaning toward Dag and the > > broader > > >>> community favouring DAG. > > >>> > > >>> It's a nice reminder that Airflow serves two audiences: Contributors > > >>> leaning toward a cleaner, more readable style, and the wider > community > > >>> still attached to the familiar "DAG" identity. Both are valid, and > it's > > >>> interesting to see how the project's voice is shifting as we grow. > > >>> > > >>> *Next Steps* > > >>> Since we'd already started shifting documentation toward Dag when it > > >> seemed > > >>> to be the general preference, the vote results essentially confirms > > that > > >>> direction. > > >>> > > >>> We'll continue using Dag in docs going forward, keeping DAG only when > > >>> referring to the class/import itself. No changes are needed for > > existing > > >>> references unless a doc is being actively updated. > > >>> > > >>> If anybody would like to call a separate vote to create a Dag alias > for > > >>> DAG, they are more than welcome to. I don't think the results of this > > >> vote > > >>> should preclude us from doing that at a later date if the community > > >> agrees. > > >>> If anyone has strong objections or follow-ups, please share them by > EOD > > >>> Friday, otherwise we'll consider this settled. > > >>> > > >>> Constance > > >>> > > >>> -- > > >>> > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > >
