You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@incubator.apache.org by Marvin Humphrey <ma...@rectangular.com> on 2012/05/27 06:44:08 UTC
Open enrollment
On Sat, May 26, 2012 at 3:02 PM, Benson Margulies <bi...@gmail.com> wrote:
> I'll see Jukka one and raise him one. I have advised potential
> podlings to be very conservative with their initial list, and keep
> some potential contributors in their collective back pocket. This
> gives them a ready-made source of community growth, which is typically
> the scarcest and most precious commodity to a podling.
+1
I'm less seasoned than others here who have done a lot of Mentoring, but my
impression is that the act of identifying, nominating and voting in a new
committer or PPMC member is a valuable experience for a fledgling community in
and of itself -- going through the process seems to be beneficial, not just in
terms of securing a new recruit and boosting their morale, but for those doing
the recruiting as they debate and become comfortable with granting privileges
to new contributors.
Adding full PPMC members during the informal "open enrollment" period prior to
the VOTE on entering incubation limits the number of times a PPMC gets to go
through this invigorating experience. Perhaps that implies that the Incubator
should actively discourage open enrollment!
On the other hand, open enrollment can be a useful recruiting tool. It was
absolutely vital for AOO (whose circumstances were unique) but it has been
valuable for other projects as well -- you don't want to squander any of the
initial excitement and publicity a new proposal generates.
I believe there may be a best-of-both-worlds solution:
* Incubation proposals should have separate sections for "Initial
Committers" and "Initial PPMC Members".
* There should be text in the proposal encouraging people to add
themselves to the list of "Initial Committers" and to introduce themselves
on general@incubator -- but no such text regarding the list of "Initial
PPMC members".
The "open enrollment" period has historically been controversial -- Crunch is
not the first project to wrestle with it. Under this proposal, we avoid
rushing not-yet-podlings into granting strangers who have not yet demonstrated
merit a governance stake, but ease them into the essential Apache process of
expanding and refreshing their community as soon as the possible.
Signing on as a committer affords anyone who joins during open enrollment the
convenience of CTR (for projects that use that process), which ought to
provide sufficient incentive to keep people joining up -- yet it also gives
the podling the opportunity to vote people into the PPMC early and often. And
if committers who demonstrate merit are *not* brought into the PPMC, the
podling's Mentors and the IPMC should take notice.
Marvin Humphrey
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Ross Gardler <rg...@opendirective.com>.
On 28 May 2012 09:30, Bertrand Delacretaz <bd...@apache.org> wrote:
> On Mon, May 28, 2012 at 12:06 AM, Josh Wills <jw...@cloudera.com> wrote:
>> ...At a minimum, I think it would be wise for the incubator documentation
>> to tell new projects to announce a policy as part of their proposal,
>> so that others do not make the same mistake. That announcement could
>> be anything from "The initial committers will be made up exclusively
>> from existing contributors" to "We would like to consider potential
>> additional committers on a case-by-case basis" to "Anyone from the
>> Apache community is welcome to join as an initial committer." Then
>> there could be explicit discussion on the thread about the pros and
>> cons of the project's choice before the proposal went to a vote....
>
> I like the idea of having an incubation proposal spell out if and how
> the proposers intend to accept new initial committers. No need to make
> this complicated, but stating that clearly is good.
+1 - less rules, more general guidelines that podlings can adapt to
their specific circumstances and needs.
Ross
>
> -Bertrand
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
--
Ross Gardler (@rgardler)
Programme Leader (Open Development)
OpenDirective http://opendirective.com
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Bertrand Delacretaz <bd...@apache.org>.
On Mon, May 28, 2012 at 12:06 AM, Josh Wills <jw...@cloudera.com> wrote:
> ...At a minimum, I think it would be wise for the incubator documentation
> to tell new projects to announce a policy as part of their proposal,
> so that others do not make the same mistake. That announcement could
> be anything from "The initial committers will be made up exclusively
> from existing contributors" to "We would like to consider potential
> additional committers on a case-by-case basis" to "Anyone from the
> Apache community is welcome to join as an initial committer." Then
> there could be explicit discussion on the thread about the pros and
> cons of the project's choice before the proposal went to a vote....
I like the idea of having an incubation proposal spell out if and how
the proposers intend to accept new initial committers. No need to make
this complicated, but stating that clearly is good.
-Bertrand
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Benson Margulies <bi...@gmail.com>.
I know that I, and I think that Marvin, got a bit distracted by their
exact choice of wording. I join him in agreeing with you that the
people proposing a podling should be entitled, no, encouraged, to
control their list of initial collaborators. Also note the other email
in this thread about wanting a consistent policy from any particular
podling.
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Ross Gardler <rg...@opendirective.com>.
On 28 May 2012 16:56, Marvin Humphrey <ma...@rectangular.com> wrote:
> On Mon, May 28, 2012 at 5:34 AM, Ross Gardler
> <rg...@opendirective.com> wrote:
>>> However, I am uncomfortable with the exclusivity of language like "The initial
>>> committers will be made up exclusively from existing contributors".
>>
>> Why? Specifically, imagine a new project coming into the incubator.
>> They have a vibrant and healthy community. They already operate
>> according to the Apache Way. They want to limit initial contributors
>> to those who have already demonstrated merit in the project. Their
>> justification is that everyone who is currently a committer earned
>> that right through hard work and they don't want to undermine that by
>> giving Joe Blogs the power of veto over critical decisions until they
>> too have proven themselves.
>>
>> In such circumstances why, as an IPMC member, would you object to this
>> approach?
>
> Hmm, there must have been a serious failure of communication between us if
> think I'd object to that. :\
OK, it's just that the comment from you quoted above ("I am
uncomfortable with the exclusivity of language like "The initial
committers will be made up exclusively from existing contributors".)
seemed pretty final. If that was not the intention then all is good.
I'm certainly in agreement that open enrollment should be an option
not a requirement.
Ross
Ross
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Marvin Humphrey <ma...@rectangular.com>.
On Mon, May 28, 2012 at 5:34 AM, Ross Gardler
<rg...@opendirective.com> wrote:
>> However, I am uncomfortable with the exclusivity of language like "The initial
>> committers will be made up exclusively from existing contributors".
>
> Why? Specifically, imagine a new project coming into the incubator.
> They have a vibrant and healthy community. They already operate
> according to the Apache Way. They want to limit initial contributors
> to those who have already demonstrated merit in the project. Their
> justification is that everyone who is currently a committer earned
> that right through hard work and they don't want to undermine that by
> giving Joe Blogs the power of veto over critical decisions until they
> too have proven themselves.
>
> In such circumstances why, as an IPMC member, would you object to this
> approach?
Hmm, there must have been a serious failure of communication between us if
think I'd object to that. :\
So long as something actually gets done to curtail open enrollment and we
aren't still talking about how "piling on" is a problem in 2018, my primary
objective in starting this thread will have been achieved, so I retire from
the field.
Marvin Humphrey
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Ross Gardler <rg...@opendirective.com>.
On 28 May 2012 13:22, Marvin Humphrey <ma...@rectangular.com> wrote:
...
>> At a minimum, I think it would be wise for the incubator documentation
>> to tell new projects to announce a policy as part of their proposal,
>> so that others do not make the same mistake. That announcement could
>> be anything from "The initial committers will be made up exclusively
>> from existing contributors" to "We would like to consider potential
>> additional committers on a case-by-case basis" to "Anyone from the
>> Apache community is welcome to join as an initial committer." Then
>> there could be explicit discussion on the thread about the pros and
>> cons of the project's choice before the proposal went to a vote.
>
> I support the idea of proposal authors taking charge of destiny and overriding
> the flawed participation policies in the Incubator's documentation.
>
> However, I am uncomfortable with the exclusivity of language like "The initial
> committers will be made up exclusively from existing contributors".
Why? Specifically, imagine a new project coming into the incubator.
They have a vibrant and healthy community. They already operate
according to the Apache Way. They want to limit initial contributors
to those who have already demonstrated merit in the project. Their
justification is that everyone who is currently a committer earned
that right through hard work and they don't want to undermine that by
giving Joe Blogs the power of veto over critical decisions until they
too have proven themselves.
In such circumstances why, as an IPMC member, would you object to this
approach? Is it that in *some* circumstances you would object to
exclusivity whilst in others you would find it acceptable? If that is
the case then shouldn't it be up to champions to guide the proposal
wording?
Ross
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Marvin Humphrey <ma...@rectangular.com>.
Hi Josh,
On Sun, May 27, 2012 at 3:06 PM, Josh Wills <jw...@cloudera.com> wrote:
> I think that not making an explicit policy announcement was a mistake on our
> part, because the interpretation was ambiguous and that led directly to an
> ugly start for the project.
Had you stated such a policy, you would be at odds with existing Incubator
participation guidelines.
http://incubator.apache.org/guides/participation.html#committer
The proposal typically contains a list of initial committers. When a
podling is bootstrapped, this list is used by the mentors to set up
initial accounts. So, one way to become a committer for a podling is to be
listed on the proposal as an initial committer.
The right way to express interest is by a post to the list with a brief
introduction. Piling onto a proposal (by adding your own name as an
initial committer) is impolite.
IMO, these guidelines are flawed -- though possibly through imprecise language
rather than intent. Encouraging people to introduce themselves on the public
list puts podling candidates in the awkward position of having to turn people
away. We've just seen how well that work out, no? :P
Skepticism of open enrollment goes back a long time. Here's ASF Director Roy
Fielding in 2006:
http://s.apache.org/VT5
In contrast, letting anyone "pile on" to a podling while it is at
the proposal stage is placing an unequal burden on a new podling
that we would never place on a full project. If the community is
not cohesive, no consensus will be possible and we effectively
hamstring the podling before it is even started.
> At a minimum, I think it would be wise for the incubator documentation
> to tell new projects to announce a policy as part of their proposal,
> so that others do not make the same mistake. That announcement could
> be anything from "The initial committers will be made up exclusively
> from existing contributors" to "We would like to consider potential
> additional committers on a case-by-case basis" to "Anyone from the
> Apache community is welcome to join as an initial committer." Then
> there could be explicit discussion on the thread about the pros and
> cons of the project's choice before the proposal went to a vote.
I support the idea of proposal authors taking charge of destiny and overriding
the flawed participation policies in the Incubator's documentation.
However, I am uncomfortable with the exclusivity of language like "The initial
committers will be made up exclusively from existing contributors".
Marvin Humphrey
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Josh Wills <jw...@cloudera.com>.
On Sun, May 27, 2012 at 12:08 AM, Jakob Homan <jg...@gmail.com> wrote:
>> The "open enrollment" period has historically been controversial -- Crunch is
>> not the first project to wrestle with it.
>
> Just to re-iterate, the issue with Crunch was not whether or not that
> group decided to have an open enrollment or not. The issue was that
> the announced policy to not have one was selectively ignored for one
> volunteer versus another. After announcing its intention to go with a
> closed enrollment, an exception was made:
>> the consensus was that your background is uniquely valuable
>> to the project, and that we would like to have you with us as an
>> initial committer.
> which is a fancy way of saying 'we're going to make an exception to
> our own policy just for your case' - a pretty bad foot for a
> merit-based effort to get off on.
The text of the entire email that Jakob is quoting from:
"Thank you Vinod. I wasn't sure of the right protocol for this sort of
thing, as my expectation was that the initial committers would be
drawn from the people who had contributed to Crunch already. This
thread from when S4 entered the incubator was particularly
illuminating:
http://markmail.org/message/aw54w4mhg4zfegpn
After talking it over with my co-submitters, the consensus was that
your background is uniquely valuable to the project, and that we would
like to have you with us as an initial committer."
There was not an email before that one that announced a policy (or
even an opinion) with respect to open or closed enrollment, and the
reference to the S4 thread only indicated that it was "illuminating,"
not that it was policy. I think that not making an explicit policy
announcement was a mistake on our part, because the interpretation was
ambiguous and that led directly to an ugly start for the project.
At a minimum, I think it would be wise for the incubator documentation
to tell new projects to announce a policy as part of their proposal,
so that others do not make the same mistake. That announcement could
be anything from "The initial committers will be made up exclusively
from existing contributors" to "We would like to consider potential
additional committers on a case-by-case basis" to "Anyone from the
Apache community is welcome to join as an initial committer." Then
there could be explicit discussion on the thread about the pros and
cons of the project's choice before the proposal went to a vote.
>
> Open enrollment, closed enrollment or the hybrid you're suggesting all
> can (or not) work fine because they're all fair rules for the new
> group to build on.
> -jakob
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
--
Director of Data Science
Cloudera
Twitter: @josh_wills
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Jakob Homan <jg...@gmail.com>.
> The "open enrollment" period has historically been controversial -- Crunch is
> not the first project to wrestle with it.
Just to re-iterate, the issue with Crunch was not whether or not that
group decided to have an open enrollment or not. The issue was that
the announced policy to not have one was selectively ignored for one
volunteer versus another. After announcing its intention to go with a
closed enrollment, an exception was made:
> the consensus was that your background is uniquely valuable
> to the project, and that we would like to have you with us as an
> initial committer.
which is a fancy way of saying 'we're going to make an exception to
our own policy just for your case' - a pretty bad foot for a
merit-based effort to get off on.
Open enrollment, closed enrollment or the hybrid you're suggesting all
can (or not) work fine because they're all fair rules for the new
group to build on.
-jakob
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Benson Margulies <bi...@gmail.com>.
Ross,
I can see why my 'sandbag' approach makes you uncomfortable. I've
suggested it once or twice when a proposed podling had a lot of
interested parties already involved. This is a two-edged situation. On
the one hand, instant size and diversity. On the other hand, that may
represent the pool of participants for some time to come, leaving the
podling a bit stalled when they have done everything to deserve
graduation except add people. Overdone, this amounts to cheating.
"George is thinking about contributing. Great, let's have him walk the
process and get added, instead of adding him to the initial list."
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Marvin Humphrey <ma...@rectangular.com>.
On Sun, May 27, 2012 at 10:09 AM, Ross Gardler
<rg...@opendirective.com> wrote:
> On May 27, 2012 1:55 PM, "Marvin Humphrey" <ma...@rectangular.com> wrote:
>> OK, then how about a place for people to sign up for the podling dev list
>> in advance?
>
> Sure, but to what end?
To welcome people on board, rather than exclude them with explicit language
in the proposal limiting participation.
> If the are interested enough they will sign up whether they signed the wiki
> page or not. Having said that there is no harm and if projects and champions
> see benefit then fair enough.
Here's a stab at some language for the proposal template at
<http://incubator.apache.org/guides/proposal.html>:
== Developer mailing list signup ==
Anyone who anticipates that they would like to join our community is
welcome to sign up in advance for our dev list (and encouraged to
introduce themselves with an email to general@incubator.apache.org)
||'''Name''' ||'''Email''' ||
|| Jane Doe || jane@example.com ||
Marvin Humphrey
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Ross Gardler <rg...@opendirective.com>.
On May 27, 2012 1:55 PM, "Marvin Humphrey" <ma...@rectangular.com> wrote:
>
> On Sun, May 27, 2012 at 12:57 AM, Ross Gardler
> <rg...@opendirective.com> wrote:
>
> >> * Incubation proposals should have separate sections for "Initial
> >> Committers" and "Initial PPMC Members".
> >
> > Too much hierarchy, the ASF is flat. This is hard to understand if we
> > introduce layers to incubation.
>
> Well, now *I'm* confused. :) The distinction between committer and PMC
member
> exists at the ASF. Some podlings choose to unify the roles of committer
and
> PPMC member, others do not -- same as with TLPs, some of which unify the
roles
> of committer and PMC Member while others do not.
Yes, and in the incubator the separation is the IPMC.
> My guess is that you believe the two roles ought to be unified
Depends on the project, entry into the incubator is not, IMHO, the time to
consider such a complex issue.
> > Why can't they just be contributors needing to earn merit just as they
> > would in a TLP?
>
> OK, then how about a place for people to sign up for the podling dev list
in
> advance?
Sure, but to what end? If the are interested enough they will sign up
whether they signed the wiki page or not. Having said that there is no harm
and if projects and champions see benefit then fair enough.
Ross
>
> In my opinion, there should continue to be a way for new people to get
> involved during the proposal phase. The reason is simple: the less time
your
> signup sheet is out there, the fewer names you'll collect.
>
> Marvin Humphrey
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
Re: Open enrollment
Posted by Marvin Humphrey <ma...@rectangular.com>.
On Sun, May 27, 2012 at 12:57 AM, Ross Gardler
<rg...@opendirective.com> wrote:
>> * Incubation proposals should have separate sections for "Initial
>> Committers" and "Initial PPMC Members".
>
> Too much hierarchy, the ASF is flat. This is hard to understand if we
> introduce layers to incubation.
Well, now *I'm* confused. :) The distinction between committer and PMC member
exists at the ASF. Some podlings choose to unify the roles of committer and
PPMC member, others do not -- same as with TLPs, some of which unify the roles
of committer and PMC Member while others do not.
My guess is that you believe the two roles ought to be unified, a position
which Chris Mattmann argued passionately for and at great length during Lucy's
incubation. I remain unpersuaded, but as I don't want to step in that
acrimonious debate, let's sidestep...
> Why can't they just be contributors needing to earn merit just as they
> would in a TLP?
OK, then how about a place for people to sign up for the podling dev list in
advance?
In my opinion, there should continue to be a way for new people to get
involved during the proposal phase. The reason is simple: the less time your
signup sheet is out there, the fewer names you'll collect.
Marvin Humphrey
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Open enrollment
Posted by Ross Gardler <rg...@opendirective.com>.
Sent from my mobile device, please forgive errors and brevity.
On May 27, 2012 5:44 AM, "Marvin Humphrey" <ma...@rectangular.com> wrote:
>
> On Sat, May 26, 2012 at 3:02 PM, Benson Margulies <bi...@gmail.com>
wrote:
> > I'll see Jukka one and raise him one. I have advised potential
> > podlings to be very conservative with their initial list, and keep
> > some potential contributors in their collective back pocket. This
> > gives them a ready-made source of community growth, which is typically
> > the scarcest and most precious commodity to a podling.
>
> +1
>
> I'm less seasoned than others here who have done a lot of Mentoring, but
my
> impression is that the act of identifying, nominating and voting in a new
> committer or PPMC member is a valuable experience for a fledgling
community in
> and of itself -- going through the process seems to be beneficial, not
just in
> terms of securing a new recruit and boosting their morale, but for those
doing
> the recruiting as they debate and become comfortable with granting
privileges
> to new contributors.
>
on almost all of the podlings I've worked with a mentor has prompted the
first nomination. In most a discussion of "is it too early" results.
Keeping known people out of the initial contributor list
In order "to keep some potential contributors in their collective back
pocket" seems wrong to me. Either someone has merit at proposal stage (put
them in) or they don't (show the community how to build and recognise
merit). I don't like even mentors being given commit status without merit
(remember merit is not transferable).
I'm not sure where this open enrollment thing came from, but I've never
liked like it. It made sense on AOO since there was a need to be open to a
previous fork. This is not the case in most other projects.
>
> I believe there may be a best-of-both-worlds solution:
>
> * Incubation proposals should have separate sections for "Initial
> Committers" and "Initial PPMC Members".
Too much hierarchy, the ASF is flat. This is hard to understand if we
introduce layers to incubation.
> * There should be text in the proposal encouraging people to add
> themselves to the list of "Initial Committers" and to introduce
themselves
> on general@incubator -- but no such text regarding the list of "Initial
> PPMC members".
Why can't they just be contributors needing to earn merit just as they
would in a TLP?
If a new project wants to have open enrollment for some reason then the
champion should advise them on a case by case basis.
Ross
Ross