incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marvin Humphrey <mar...@rectangular.com>
Subject Re: Incubation capacity
Date Mon, 12 Oct 2015 18:50:03 GMT
On Mon, Oct 12, 2015 at 8:08 AM, Jukka Zitting <jukka@zitting.name> wrote:

> It sounds like ruminations about the Incubator are on the increase again,

I hope that we can make use of some of this bursting energy and channel it
into incremental improvements.

The Incubator is a stable platform, and it has been functioning well by
historical terms, and with blessedly low drama compared to a few years ago.
My impression is that frustration with the institutional resistance of
Incubator to change is skewing impressions of how well it is doing its job of
incubating podlings.

> I believe the way the Incubator is organized sets an upper bound on the
> number of podlings it can effectively manage. Based on experience and
> historical data (http://incubator.apache.org/history/ *) I believe this
> limit is somewhere around 30 podlings.

I'm curious, Jukka.  Why 30?  What are the scarce resources?  And how is this
supposed degradation manifesting?

*   One single point of failure has been the Chair, especially with regards to
    the report.  However, we have addressed that to some extent by having
    other people produce the report while leaving the Chair with editorial
    prerogative.
*   You could argue that we have a shortage of shepherds since participation
    has been flagging, but I don't see that as a bad thing.  The critical
    function of the shepherds -- keeping neglected podlings from falling
    through the cracks -- is now handled by the Report Manager.  And the
    expectation that it is the shepherd's role to provide review of podling
    reports is harmful in that it discourages Mentors from commenting.
*   It used to be the case that we were perpetually short on IPMC votes for
    releases, but since late 2013 we have been using that limited capacity
    much more effectively -- thanks to improved understanding of release
    policy and improved consensus about when leniency is appropriate, we now
    cycle through fewer release candidates.

Additionally, I'll note that while we're at 43 or so podlings right now, we
have multiple podlings about to retire (Droids, Kalumet, likely Corinthia) and
others about to graduate (Kylin, Groovy).

It would probably be healthy to nudge some podlings here and there towards
graduation or retirement, but we're doing that.  Podlings which don't report
get put into the "monthly" reporting category, as are podlings whose reports
are not signed-off by their Mentors -- and they are not removed until they
provide a proper signed-off report.  Recently this has triggered action on
Droids, Kalumet, and Wave.  Before that, NPanday and others.

So... I can see an argument that some individual Mentors might sometimes
stretch themselves too thin by taking on multiple projects, but I don't yet
grok how the Incubator's central institutions are having capacity problems.

Marvin Humphrey

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message