[sakai2-tcc] Thoughts on 2.9.1 schedule

[Neal Caidin]

Okay, I'm open to discussion on this.

Maybe I'm being too risk averse?  Maybe we don't need as much QA on 2.9.1-RC01 so perhaps we could get it out in a week or two instead of several weeks?

My desire is to have a quality release and a realistic schedule as we get near the "end game".  I don't mind having aggresive or stretch schedules on larger efforts.

I also need to make sure that the resources are committed to the schedule. I'm not sure how y'all typically account for the potential impact of local issues on schedules. 

If several TCC members are available to have a quick teleconference on Friday, I'm happy to discuss. I don't want to give the impression that I want the release to drag on. I don't. I just want to do what I can to insure it is quality and to communicate a realistic schedule. 

Any takers for a quick discussion?  Maybe 10 am Eastern, 3 pm GMT? 

Thanks,
Neal


On Jan 17, 2013, at 4:01 PM, Noah Botimer <botimer at umich.edu> wrote:

> TL;DR - we can go faster unless we screwed up in ways we want to avoid in the future. Another month for something we feel is somewhere around a month late is a bad sign.
> 
> ------
> 
> Just a gentle reminder... We have always tried to keep the .x branches releasable at any time. In fact, this very principle has enabled people to run from the branch rather than releases (which, paradoxically, has reduced our ambition for releases and hurt our coordination).
> 
> If this has changed somehow in terms of substance (untested changes in branches) and not just a matter of scoping what seems to be a reasonable and compelling package of changes, we have a new and different problem on our hands -- that our implicit (and often made explicit) contract with the community has changed without sufficient clarity.
> 
> I know how difficult it is to resist the "just one more fix/day/week" urge -- I am quite guilty, myself (even for this release). But we have, I think, reached critical mass for a release and any further scope only makes the load heavier to carry across the line. If we need to make a call for massive, lengthy QA, we've been merging stuff before it's ready. If we don't know what's in 2.9.x, we've been merging stuff without proper care. If there are more issues than we can test (plus a little broad sanity checking) in a day or two, we are way overdue on release and are falling victim to the time-scope-difficulty exponent.
> 
> Please believe that I am not trying to chastise anyone -- that's not helpful. I am just trying to nudge us toward actualizing the goals we have been trying to reach for quite some time (regular, understood, communicated maintenance releases that are relatively easy to execute and adopt).
> 
> I also recognize that our actual release and upgrade mechanics are still not trivial. This is not a call for weekly releases. It is just a restatement of the lovely agile axiom: if something hurts, do it more often [1]. Lots of good comes from that simple change.
> 
> And, finally... Thanks to everyone who has been working hard. I'm not asking you to work harder. I'm asking us to take a look at our patterns and improve our release process health by exercising the activities we want to become habits.
> 
> Thanks,
> -Noah
> 
> http://martinfowler.com/bliki/FrequencyReducesDifficulty.html
> 
> On Jan 17, 2013, at 3:28 PM, Neal Caidin wrote:
> 
>> Hi TCC,
>> 
>> Before I publish to the broader community. I'm thinking that 2.9.1 will be out between mid-February and late-February. It of course is dependent on Matt J.'s schedule and local workload, amount of QA testing we get, availability of QA coordination (normally this has been me), and if any blocker issues come up necessitating an RC02. I think we have a 50/50 chance of only needing an RC01 since 2.9.x has had some pretty good testing against it. We also have a few security issues which I'm hoping we can get tested and merged, but if we don't, we might want to proceed anyway and target them for 2.9.2 . 
>> 
>> Details
>> ----------------
>> Since Monday is a holiday, the soonest we can get a 2.9.1-RC01 out is next Thursday morning (or the night before), Jan 24. The build could start on Tuesday and usually takes two days to complete. 
>> 
>> Scenario 1 - optimistic scenario - assuming an RC01 only and sufficient QA coordination and activity
>> ------------------------------------------------------------------------------------------------------------------
>> Earliest start :
>> Tuesday - Jan 22 - RC01 build starts
>> Thursday - Jan 24 - QA testing begins
>> Friday - Feb1 - QA testing complete
>> Monday - Feb 4 - Final build commences - GA
>> Wednesday - Feb 6 - final build done for GA (general availability)
>> Thurs/Fri - Final validation testing against GA, and documentation
>> Friday, Feb 8 - announce official release
>> 
>> If we don't have sufficient QA coordination and testing, you could add another week on. If we need an RC02 then add another 2 weeks (or at least week and a half) so we can get things fixed, a new build, and QA testing. That would put us at March 1 (but we can tweak that to Feb 28).
>> 
>> Sound reasonable?  
>> 
>> I would announce to community that current target for 2.9.1 looks like mid to late February.
>> 
>> Cheers,
>> 
>> 
>> Neal Caidin
>> 
>> Sakai CLE Community Coordinator
>> nealcaidin at sakaifoundation.org
>> Skype: nealkdin
>> AIM: ncaidin at aol.com
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> sakai2-tcc mailing list
>> sakai2-tcc at collab.sakaiproject.org
>> http://collab.sakaiproject.org/mailman/listinfo/sakai2-tcc
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai2-tcc/attachments/20130117/6d2035f6/attachment-0001.html