incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Niclas Hedhman <nic...@hedhman.org>
Subject Re: Incubator report sign-off
Date Tue, 23 Dec 2014 07:54:21 GMT
On Tue, Dec 23, 2014 at 12:42 AM, Roman Shaposhnik <rvs@apache.org> wrote:

>     1. get rid of IPMC altogether and move to the pTLP model
>     2. make this a poddling issue: if a poddling fails to hunt down ALL
>         the mentors for a sign-off -- reject its report
>     3. patch the current process with starting to drop the mentors from
>         the project who don't sign off. This will essentially serve
>         as a heartbeat for mentors (now, in my opinion it'll quickly
>        deteriorate into mindless tick-offs -- but at least it does not
harm).

There is a fairly long tradition in ASF to not make any irreversible steps,
so "get rid of the IPMC" isn't an option until "pTLP model" has been tried,
tested and evaluated. The "move to" is in the wrong direction...

Doug even suggested "Put proposals in front of the Board, and see what
happens" (my words, not his). I would imagine years from "first pTLP" to
"finally get rid of IPMC", so if #2 is not an option, neither is the first
5 words of #1.

I always felt that the "many mentors" idea was an anti-pattern, as sense of
responsibility quickly declines for the individual mentor -- "the others
can take care of this"-attitude.

While Provisional TLP proposals are tested, I would favor;

   a. single mentor, the "always present person" who is "first person to
ask" and "with enough karma", the friendly neighborhood convenience store,
the oracle of Delphi. I would even like to see him/her be the first PMC
Chair on graduation, to raise the expectation of responsibility.

   b. incubating projects don't make ASF releases, hence the IPMC doesn't
need to vote the release, just like the Board doesn't vote on a TLP's
release. Let the ppmc (with Mentor help) handle it all in-house. They MAY
announce releases on general@, but MUST bring it up in the reports.

   c. podling acts identically to a TLP, except IPMC acts "as Board"

   d. hence IPMC assigns a Shepard (possibly rotating), who checks report
sent, check the mentor is active and that podling is "healthy".

This better matches the ASF modus operandi, less changes on graduation and
solves the AWOL mentor problem. But it also creates new ones; fewer
volunteers for mentor role, transitional pain from current system, and
probably much else I haven't thought of.


Cheers
--
Niclas Hedhman, Software Developer
http://www.qi4j.org - New Energy for Java

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message