On 1.2 timing

From: Andrus Adamchik (andru..bjectstyle.org)
Date: Fri Mar 10 2006 - 00:59:18 EST

  • Next message: Andrus Adamchik: "Re: incubator status page"

    IMO two things are almost equally important to us and our users -
    release 1.2 ASAP and make it a high quality release. While I am happy
    with 1.2 results, I must admit that there was a miscalculation in
    trying to accomplish too many things at once that required
    fundamental stack changes. Next time we need to settle on something
    more manageable and well defined (still great new ideas pop up
    constantly and it is barely possible to ignore them or put on hold
    indefinitely).

    First regarding the suggestion to branch. We can't do that until 1.2
    code stabilizes (more on this below). Maintaining two branches
    (especially with incompatible folder structure!) is going to be a
    nightmare. -1 on that. I suggest that we initially limit the
    migration to cayenne-jpa and cayenne-jpa-tck (things that are not a
    part of 1.2) and migrate the main code around the time when 1.2RC1
    goes out.

    So where do we stand right now? We have three major new features that
    work, but may use some improvement: remote object persistence, nested
    contexts and xml encoding. (we discussed some cgen usability
    improvements with Mike in the past, but let's leave those out for now).

    I suggest that we remove two other things on the roadmap (EOF
    integration improvements and user properties in the DataMap) from the
    1.2 schedule, although I feel bad about both of them, as they were
    often requested. I'll try to organize the remaining stuff on Jira to
    something more resembling a formal road map, cutting off everything
    we could realistically reschedule.

    Regarding the schedule. I suggest that we do M12, and then B1
    (framework code freeze). Beta is usually the time to fix bugs and
    improve the Modeler and docs. Then we go RC (full code freeze). This
    is when migrating the code to Apache package names becomes
    realistically possible.

    Andrus



    This archive was generated by hypermail 2.0.0 : Fri Mar 10 2006 - 00:59:21 EST