Hi Greg,
On Fri, Jan 23, 2015 at 1:33 PM, Greg Stein <gst...@gmail.com> wrote:
> There is nothing stopping the IPMC from designating certain Mentors as
> shepherds for their podlings.
Having volunteers step forward as dedicated shepherds for individual
podlings would be helpful. On its own, though, it is not sufficient,
because Incubator shepherds are not as reliable as Board members. What
happens when the dedicated shepherd goes missing? Podlings will start
falling through the cracks again.
An additional mechanism needs to be in place to ensure that no report goes
unreviewed. For instance:
1. The Incubator doesn't file its report until each and every podling
report has been reviewed by either a shepherd or a freelance IPMC
member filling in.
2. Podling reports which have not been reviewed by a shepherd are omitted
from the aggregate report and the podling is required to report again
next month.
Starting this month, the Incubator has instituted something similar to the
second option: podling reports where not a single Mentor has signed
off get rejected and the podling is required to report again next
month[1]. There was criticism that this mechanism punishes a podling for
the sins of its Mentors, but the intended result was achieved: every
podling report got signed off. (Besides, in many cases the podling is at
fault for filing at the last minute and leaving too small a window for
Mentor signoff.)
With signoff required, Mentors assume the essential functionality of
shepherds, and the value added by the titular "shepherds" is limited to
cross-cutting feedback. I maintain that there are better ways to provide
such feedback.
> That activity
> wasn't happening in the past, so the shepherds were filling in.
Shepherd participation has fallen too low to keep podlings from getting
lost -- it's now below 50%. What has kept distressed podlings like
NPanday from falling off the IPMC's radar screen, for the last year and a
half, has been the Report Manager putting podlings who don't file reports
into "monthly" reporting. It's not perfect, but it's *way* less work and
more reliable than shepherds.
Maybe the Incubator should strike that task from the Report Manager's
runbook and start losing track of podlings again? Because I feel like we
designed a better system and nobody noticed.
Marvin Humphrey
[1] This is related but not linked to the list of not-signing-off Mentors
which the Board has chosen to remove from this month's report. I've
remained silent about that up till now out of deference to those IPMC
members who are working hard to address issues of Mentor
accountability, but I support the Board's decision.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org
