[Dev] Last call for maintainer before force "Accept" "Submit" (slow progress)

Carsten Haitzler (The Rasterman) tizen at rasterman.com
Fri Apr 4 00:03:03 GMT 2014


On Thu, 03 Apr 2014 19:53:04 +0200 Zofia Abramowska <z.abramowska at samsung.com>
said:

> 
> 
> > -----Original Message-----
> > From: Carsten Haitzler (The Rasterman) [mailto:tizen at rasterman.com]
> > Sent: Thursday, April 03, 2014 4:38 PM
> > To: Zofia Abramowska
> > Cc: dev at lists.tizen.org
> > Subject: Re: [Dev] Last call for maintainer before force "Accept"
> > "Submit" (slow progress)
> > 
> > On Thu, 03 Apr 2014 14:53:22 +0200 Zofia Abramowska
> > <z.abramowska at samsung.com>
> > said:
> > 
> > >
> > > > -----Original Message-----
> > > > From: dev-bounces at lists.tizen.org
> > > > [mailto:dev-bounces at lists.tizen.org]
> > > > On Behalf Of Carsten Haitzler (The Rasterman)
> > > > Sent: Saturday, March 29, 2014 3:20 AM
> > > > To: Dominig ar Foll
> > > > Cc: dev at lists.tizen.org
> > > > Subject: Re: [Dev] Last call for maintainer before force "Accept"
> > > > "Submit" (slow progress)
> > > >
> > > > On Fri, 28 Mar 2014 18:32:14 +0100 Dominig ar Foll
> > > > <dominig.arfoll at fridu.net>
> > > > said:
> > > >
> > > > > Hello,
> > > > >
> > > > > After 3 days we have only 3 pending patches which have
> > progressed.
> > > > > Still 19 to go.
> > > > > See attached file.
> > > > > Maintainers, please step up.
> > > > >
> > > > > Regards
> > > >
> > > > i can tell you what the problem likely is...
> > > >
> > > > the people you are trying to ping are not even subscribed here.
> > most
> > > > developers who work on stuff that's related to tizen around me
> > > > simply are not on any tizen mailing lists. also the email system
> > > > they are meant to use (webmail) doesn't have filtering, deletes all
> > > > mail after 2 weeks, and has no threading abilities...
> > > >
> > > > they just never know you said anything... :) most likely. so you
> > are
> > > > shouting at an empty room.
> > > >
> > > > i have a suggestion to cure this.
> > > >
> > > > 1. we require anyone doing any tizen development (especially anyone
> > > > who has to review or approve something) to be subscribed to this
> > > > mailing list, and must read it actively. if people want to be
> > > > involved they need to be contactable and actually respond. they
> > need
> > > > to know about general discussions etc.
> > > > 2. given #1, we send out a ping here to EVERYONE that they must
> > > > respond to on this mailing list. ensure the mail is not just a few
> > > > lines (make it 50 or 100 lines long) so you see if they READ their
> > > > email or not (eg if it's more than 2 lines they skip it).
> > > > 3. wait 2 weeks.
> > > > 4. everyone who has not responded after 2 weeks gets removed from
> > > > having any review/maintainer roles. replace those people with some
> > > > generic maintainer person who does respond, so things can move
> > along.
> > > >
> > > > i've been following this "someone please look at my review" thing
> > > > for a while now. there are pings here and it regularly fails to get
> > > > attention. people are not doing their jobs it seems. they don't
> > take
> > > > their responsibility seriously enough to respond within a
> > reasonable
> > > > time. i see no alternative but to be harsh at this point. also
> > > > publish the list of all people removed so there is a bit of public
> > > > naming and shaming. it's time to learn the hard way.
> > > >
> > > > --
> > > > Carsten Haitzler (The Rasterman) <tizen at rasterman.com>
> > > > _______________________________________________
> > > > Dev mailing list
> > > > Dev at lists.tizen.org
> > > > https://lists.tizen.org/listinfo/dev
> > >
> > > People sometimes go on vacation/holidays, so this "2 weeks deadline"
> > > does not really speak to me well, especially when we're talking about
> > > naming and shaming...
> > 
> > we are getting complaints on no action after just a few days (dominig
> > complained after 3 days).
> > 
> > 2 weeks is pretty generous to just respond to a ping. as a comparison -
> > i just removed commit access from a bunch of people who haven't
> > committed for more than 6 months (to an upstream project) -- went over
> > logs to see who has committed. they had about 3 days to respond to ask
> > to get off the "purge" list.
> > everyone who wanted off and had a reason responded within that time. 3
> > days worked well there.
> > 
> > the point of a ping is to find out which people are not responding
> > quickly and
> > 2 weeks is generous as a margin of "respond quickly". many may newver
> > read this mailing list or are even subscribed. that means they are
> > unreachable except directly and personally and a project cant be
> > sustained if everything has to be a direct personal contact.
> > 
> > > When it comes to maintainers, maybe there should be some mechanism to
> > > temporarily pass the rights of maintainer to someone else, when
> > he/she
> > > is not able to fulfill his/her role.
> > 
> > that is not the problem. it's not a temporary thing. it seems to be
> > continual where things are not being reviewed for many weeks or even
> > months on end.
> 
> I am aware it's not a temporary thing. I'm only suggesting, that in some
> specific cases removing from maintainers after 2 week silence might be
> unfair.
> 
> > so if someone happened to go on holiday for 2 weeks,
> > then they can catch up when they come back.
> 
> Yes, exactly. But if they will miss this ping mail your solution proposes
> removing them from maintainers. How are they suppose to catch on then? (Of
> course they can ask for these rights again, but I'd like to avoid something
> like this).

they can ask for they rights back - yes, but after 2 weeks of not even
responding to a ping... so how long is reasonable? 1 month? 2 months? 3 months?
6 months? you know that a mail just scrolls off most peoples inboxes after a
few days and is barely looked at again. some organizations force mail to be
deleted after some period (eg 2 weeks), so it is impossible to catch up and
respond after a period anyway. so how long is it reasonable for someone to wait
for some response on a review? how long is it reasonable to wait for the
ability to communicate with the people who, by design of the organization, are
REQUIRED to respond just for things to move forward at all?

i would argue the gerrit review system doesn't work for tizen. it's overhead
that isn't functioning. the os can't move forward at all without review, unlike
other groups which may allow people to have direct commit access if they are
trusted.

i would argue it is MORE unfair for people to have their work blocked
indefinitely or for long periods, when they are trying to actively get
something done, than it is unfair to shame and remove people NOT responding.
like all things - it's balancing out what is more fair or unfair. if you want
to be fair to everyone then you can place no time limit on a response, which
means the project is allowed to stall indefinitely, and that is perfectly ok. i
am sure people looking at deadlines don't agree with you on that. :)

> > that is not the case here.
> > when it's your job to respond to review requests in a timely fashion,
> > then you certainly respond within 2 weeks or ensure someone is
> > responding for you if you will be away for an extended time (and very
> > few people will be, as most people involved in review here are in
> > countries where you can barely even take 2 weeks off in an entire
> > year).
> > 
> 
> I don't want to discuss countries, etc. Someone can be as well absent
> because of health problems for few months.

then they are not doing their job. if your manager is sick for 3 months but you
need his approval to let's say go on a necessary business trip, then he is
stalling the organization. someone else should be appointed to take over, and
he can ask for his approval powers back when he gets back. if someone not
responding to a request stops functioning of a group, then it's a problem, fair
or unfair, some weeding out has to be done.

> Don't get me wrong, I do understand it's a big and frustrating problem.
> 
> I totally agree the enforcement of subscription to this mailing list for all
> developers (or at least maintainers/integrators). I'm just trying to suggest
> being more careful in shaming anyone.

why not? being given the responsibility to do something doesn't seem to
motivate. a request to "get on with it" doesn't seem to get much done (on the
mailing list where everyone should be at least actively following discussions),
so then what? i think 2 weeks is reasonable. it is actually generous. it's not
asking for all reviews to be finished. it's simply asking for a "hey - are you
here?". like a roll call. if you are not there for the roll call.

-- 
Carsten Haitzler (The Rasterman) <tizen at rasterman.com>


More information about the Dev mailing list