You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@incubator.apache.org by Ralph Goers <ra...@dslextreme.com> on 2009/08/05 03:39:46 UTC
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
On Aug 4, 2009, at 2:10 AM, Jukka Zitting wrote:
> Hi,
>
> On Tue, Aug 4, 2009 at 10:10 AM, Bertrand
> Delacretaz<bd...@apache.org> wrote:
>> In such a case, the key point is, do the people who write the code
>> listen to the community?
>
> That's certainly good, but IMHO not really the key point as it doesn't
> address the case when those people lose interest. A community that
> actually writes the code is much stronger than one that just helps key
> committers do that job.
>
> The River and PDFBox podlings that I'm currently mentoring are good
> examples of projects where community input was valued and taken into
> account by the key committers, but when they no longer had
> time/interest the projects essentially stopped as there was nobody to
> continue the work. PDFBox seems to have overcome that problem now and
> River is showing some positive signs, but both cases have required
> (and still require) quite a bit of mentoring to get them going again.
>
> Besides all the other good things diversity brings it's also an
> insurance policy for the project, and that's what I think the
> Incubator should be looking for as a graduation criteria.
>
You are trying to predict the future. Good luck with that.
The rules are there because we have a belief that meeting them will
give the project the best chance to succeed. I would argue that if
this is your measure you should take a look at Logback and SLF4J. The
number of people who have commit rights is very small (essentially 1
in the case of Logback). But Ceki is a recognized expert in the field
and is passionate about logging. The odds of his abandoning the
project are about equal to that of him getting hit by a truck. But
there are severl active participants in the projects (myself included)
and many more who stop by and ask questions.
Using these projects as an example is perhaps not the best from a
community perspective because Ceki has no intention of running them
like Apache projects. But even if he did, by these standards the
projects might never make it out of the incubator. Even if those of us
who would like them had commit rights I can guarantee that 95% of the
commits would still be Ceki's.
From my perspective we should be evaluating projects based on whether
they are building a healthy Apache community where we have sufficient
belief that the project will be able to sustain itself without further
mentoring. IMO, trying to factor in "what ifs" about what will happen
if certain committers leave, unless they have shown signs that that is
likely, is not a very good indicator of success. Over the course of
time a project is in the incubator I would expect that mentors to have
a good sense of what the level of commitment is and would use that as
part of their recommendation for graduation.
Ralph
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Niclas Hedhman <ni...@hedhman.org>.
Yes,
And here is another angle; ALL projects will die, it is just a matter of
time.
Now, knowing that we need to deal with, and can't set out with the notion
that if we think it might whither it can't graduate, then I will need to
vote against all graduations.
So, IMHO, the "might die" criteria is sooo wrong, it makes me sick. I look
at it the other way; If it graduates it might flourish. And if one podling
do so (for a while), I am Ok with a couple that didn't.
If your intentions are for projects to die sooner, then keep them in
incubation...
-- Niclas
On Aug 5, 2009 9:40 AM, "Ralph Goers" <ra...@dslextreme.com> wrote:
On Aug 4, 2009, at 2:10 AM, Jukka Zitting wrote: > Hi, > > On Tue, Aug 4,
2009 at 10:10 AM, Bertra...
You are trying to predict the future. Good luck with that.
The rules are there because we have a belief that meeting them will give the
project the best chance to succeed. I would argue that if this is your
measure you should take a look at Logback and SLF4J. The number of people
who have commit rights is very small (essentially 1 in the case of Logback).
But Ceki is a recognized expert in the field and is passionate about
logging. The odds of his abandoning the project are about equal to that of
him getting hit by a truck. But there are severl active participants in the
projects (myself included) and many more who stop by and ask questions.
Using these projects as an example is perhaps not the best from a community
perspective because Ceki has no intention of running them like Apache
projects. But even if he did, by these standards the projects might never
make it out of the incubator. Even if those of us who would like them had
commit rights I can guarantee that 95% of the commits would still be Ceki's.
>From my perspective we should be evaluating projects based on whether they
are building a healthy Apache community where we have sufficient belief that
the project will be able to sustain itself without further mentoring. IMO,
trying to factor in "what ifs" about what will happen if certain committers
leave, unless they have shown signs that that is likely, is not a very good
indicator of success. Over the course of time a project is in the incubator
I would expect that mentors to have a good sense of what the level of
commitment is and would use that as part of their recommendation for
graduation.
Ralph
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 2:44 PM, Niall Pemberton wrote:
>>
>
> OK we have a similar example here at the ASF - when Craig McC. left
> Apache Shale it slowly died - and AFAIK become the first project to
> join the Attic. So Ceki decides to become a Yak farmer in patagonia
> and maybe the same thing happens to SLF4J. I'm in two minds about
> Pivot's graduation - but I do believe that for its long term health it
> needs to get more people actively working on the code.
>
Actually, if Ceki "moved on" there is no way SLF4J or Logback would
die. They might have to move somewhere else so others could manage the
source repository, mailing lists, etc., but there are too many other
people involved in the community for it to die. That is precisely the
point I was trying to make. I'm not using those projects as good
examples of Apache projects - they aren't. But they are great examples
of why you can't really judge the health of a project by how many
people are committing code. In fact, if you look at the mailing lists
and commit stats for SLF4J & Logback vs Commons Logging & Log4j you
might be surprised at how much more active the former are than the
latter.
As for Shale, my understanding is that it was sort of a proving ground
for things that were eventually to be incorporated into JSF 2.0. That
seems to be confirmed by http://blogs.jsfcentral.com/editorsdesk/entry/shale_in_the_attic
. But Shale was a TLP, not an incubator project. Somehow the project
managed to meet the requirements to become one. I don't know the
details of how it went from a Struts subproject to a TLP and whether
that was really warranted, but the one thing you haven't mentioned was
anything about the community. The problem isn't projects with a single
(or two) committer(s) with lots of community involvement. The problem,
as this case illustrates, is a community with a single active
participant and no one to take over if they leave.
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Niall Pemberton <ni...@gmail.com>.
On Wed, Aug 5, 2009 at 2:39 AM, Ralph Goers<ra...@dslextreme.com> wrote:
>
> On Aug 4, 2009, at 2:10 AM, Jukka Zitting wrote:
>
>> Hi,
>>
>> On Tue, Aug 4, 2009 at 10:10 AM, Bertrand
>> Delacretaz<bd...@apache.org> wrote:
>>>
>>> In such a case, the key point is, do the people who write the code
>>> listen to the community?
>>
>> That's certainly good, but IMHO not really the key point as it doesn't
>> address the case when those people lose interest. A community that
>> actually writes the code is much stronger than one that just helps key
>> committers do that job.
>>
>> The River and PDFBox podlings that I'm currently mentoring are good
>> examples of projects where community input was valued and taken into
>> account by the key committers, but when they no longer had
>> time/interest the projects essentially stopped as there was nobody to
>> continue the work. PDFBox seems to have overcome that problem now and
>> River is showing some positive signs, but both cases have required
>> (and still require) quite a bit of mentoring to get them going again.
>>
>> Besides all the other good things diversity brings it's also an
>> insurance policy for the project, and that's what I think the
>> Incubator should be looking for as a graduation criteria.
>>
>
> You are trying to predict the future. Good luck with that.
>
> The rules are there because we have a belief that meeting them will give the
> project the best chance to succeed. I would argue that if this is your
> measure you should take a look at Logback and SLF4J. The number of people
> who have commit rights is very small (essentially 1 in the case of Logback).
> But Ceki is a recognized expert in the field and is passionate about
> logging. The odds of his abandoning the project are about equal to that of
> him getting hit by a truck.
OK we have a similar example here at the ASF - when Craig McC. left
Apache Shale it slowly died - and AFAIK become the first project to
join the Attic. So Ceki decides to become a Yak farmer in patagonia
and maybe the same thing happens to SLF4J. I'm in two minds about
Pivot's graduation - but I do believe that for its long term health it
needs to get more people actively working on the code.
Niall
> But there are severl active participants in the
> projects (myself included) and many more who stop by and ask questions.
>
> Using these projects as an example is perhaps not the best from a community
> perspective because Ceki has no intention of running them like Apache
> projects. But even if he did, by these standards the projects might never
> make it out of the incubator. Even if those of us who would like them had
> commit rights I can guarantee that 95% of the commits would still be Ceki's.
>
> From my perspective we should be evaluating projects based on whether they
> are building a healthy Apache community where we have sufficient belief that
> the project will be able to sustain itself without further mentoring. IMO,
> trying to factor in "what ifs" about what will happen if certain committers
> leave, unless they have shown signs that that is likely, is not a very good
> indicator of success. Over the course of time a project is in the incubator
> I would expect that mentors to have a good sense of what the level of
> commitment is and would use that as part of their recommendation for
> graduation.
>
> Ralph
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Robert Burrell Donkin <ro...@gmail.com>.
On Wed, Aug 5, 2009 at 4:32 PM, Ralph Goers<ra...@dslextreme.com> wrote:
>
> On Aug 5, 2009, at 8:21 AM, Ralph Goers wrote:
>
>>
>> On Aug 5, 2009, at 2:00 AM, Jukka Zitting wrote:
>>
>>> Hi,
>>>
>>> On Wed, Aug 5, 2009 at 3:39 AM, Ralph Goers<ra...@dslextreme.com>
<snip>
>>> That's why I measure the "three independent committers" criteria by
>>> looking at the commit log instead of the asf-authorization file. And
>>> I'm not asking much, just a few code commits in the past few months is
>>> good enough for me.
>>>
>>> That's the criteria that I held Sling against, and that's also what's
>>> currently keeping UIMA from graduating (and apparently also for over a
>>> year before I signed up to help them). If the consensus is that this
>>> is a bit too hard a requirement, then I'll be happy to bring UIMA up
>>> for graduation in the next few months.
>>>
>>
>> I'm suggesting that while that criteria makes sense for someone not
>> involved in a project, it isn't necessarily the best way to determine how
>> healthy the community is. If mentors are mentoring then we should give
>> heavy weight to their recommendation and judgment with respect to
>> graduation. If a mentor needs these kinds of statistics to tell him/her
>> whether the project is healthy then I am very concerned.
>>
>> Ralph
>
> I should also clarify. Some projects will never be ready to stand on their
> own as a TLP. Sanselan moved to Commons for just this reason. I'm not sure
> that Commons is the best home for every project in this position, but as
> part of the graduation recommendation mentors should identify this. If the
> only concern over a community is its size (but they clearly understand the
> Apache way of doing things) then graduation as a subproject makes complete
> sense to me and we should actively look for the right TLP to move it to.
in the case when there is no suitable TLP, do we need to start
proposing new commons-style TLPs?
- robert
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 8:21 AM, Ralph Goers wrote:
>
> On Aug 5, 2009, at 2:00 AM, Jukka Zitting wrote:
>
>> Hi,
>>
>> On Wed, Aug 5, 2009 at 3:39 AM, Ralph Goers<ralph.goers@dslextreme.com
>> > wrote:
>>> Using these projects as an example is perhaps not the best from a
>>> community
>>> perspective because Ceki has no intention of running them like
>>> Apache
>>> projects. But even if he did, by these standards the projects
>>> might never
>>> make it out of the incubator. Even if those of us who would like
>>> them had
>>> commit rights I can guarantee that 95% of the commits would still
>>> be Ceki's.
>>
>> I don't see it as a problem if the vast majority of commits comes
>> from
>> one person (or company) as long as the community operates normally
>> *and* there are others who won't have to start learning how to build
>> the codebase and do an "svn commit" if the key developer leaves.
>
> You won't find the second part solely from the commit log. I would
> expect mentors to be monitoring the dev list. If it is filled with
> "can you fix this?" then there is a problem. But if it also has "why
> did you do this?", "can we do this?", "I don't understand this
> commit" type messages then I expect the second concern is adequately
> addressed. But you can't determine that from raw statistics.
>
>>
>> That's why I measure the "three independent committers" criteria by
>> looking at the commit log instead of the asf-authorization file. And
>> I'm not asking much, just a few code commits in the past few months
>> is
>> good enough for me.
>>
>> That's the criteria that I held Sling against, and that's also what's
>> currently keeping UIMA from graduating (and apparently also for
>> over a
>> year before I signed up to help them). If the consensus is that this
>> is a bit too hard a requirement, then I'll be happy to bring UIMA up
>> for graduation in the next few months.
>>
>
> I'm suggesting that while that criteria makes sense for someone not
> involved in a project, it isn't necessarily the best way to
> determine how healthy the community is. If mentors are mentoring
> then we should give heavy weight to their recommendation and
> judgment with respect to graduation. If a mentor needs these kinds
> of statistics to tell him/her whether the project is healthy then I
> am very concerned.
>
> Ralph
I should also clarify. Some projects will never be ready to stand on
their own as a TLP. Sanselan moved to Commons for just this reason.
I'm not sure that Commons is the best home for every project in this
position, but as part of the graduation recommendation mentors should
identify this. If the only concern over a community is its size (but
they clearly understand the Apache way of doing things) then
graduation as a subproject makes complete sense to me and we should
actively look for the right TLP to move it to.
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 9:58 AM, Martijn Dashorst wrote:
> AFAIK Nowhere in the policy it is stated that committer activity ==
> code commits.
>
> Martijn
Absolutely right. I'd actually like to see more than 3 people with
commit privs (and I think most projects have that). As a mentor I
would be looking to see that the community is healthy - that there are
discussions about code going on and that there is a clear indication
that more than 1 or 2 people understand the code. I really don't think
it is all that important who actually performs commits, although it is
a great indicator that this is happening.
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Niclas Hedhman <ni...@hedhman.org>.
I will try to answer it;
1. I think it is an extreme question. If both disappeared tomorrow, I don't
think the will to continue exists. If it happens 12 months from now in the
Incubator, I think it could go either way. If happens 12 months after
graduation, I am pretty sure it sustains. Why do I think so? It is all about
perception. Personally, I would think twice trying to blow life into a
incubating project with problems, but not so with a TLP that I depend on.
Pivot is sexy, possibly highly controversial and competing head on with Flex
and gang. I don't think it will get many serious users until it is out of
incubator, which in turn stimulate the people working on it a lot more.
2. Greg and Todd from the start (Feb2009) very open and discussed most
things on ML and constantly asking for feedback. The early trickle of dev
questions, incl my own, often resulted in "Yeah, you are right, a great
idea..." and before I knew it has been incorporated. At least as good as
your regular TLP. Since people like stats, ~200 tickets in 6 months is IMHO
another sign of transparency.
3. Yes I think so. In fact, I suspect it will end up being an umbrella of
smaller projects using Pivot but where that shouldn't go into Pivot itself.
I think a similar evolution as Apache Felix, perhaps slightly slower growth
(due to OSGi's popularity when Felix came here). I also can't identify any
existing TLP suitable.
Cheers
Niclas
On Aug 6, 2009 10:31 AM, "Ralph Goers" <ra...@dslextreme.com> wrote:
On Aug 5, 2009, at 10:15 AM, Jukka Zitting wrote: > Hi, > > On Wed, Aug 5,
2009 at 6:58 PM, Martij...
I don't think your concerns about Pivot are baseless. I think that stating
them in terms of number of commits doesn't really frame your concern
properly.
Here is how I would express my concerns based on the messages I've read
here.
1. We know there are two committers who understand the code very well. Do
the mentors feel the rest of the community has the necessary understanding
and commitment to pick up the slack should something happen to those two
individuals?
2. Does the community participate in the direction of the project? Do the
committers do whatever they want without review or do others participate?
3. If Pivot were given permission to graduate would the project really have
the momentum it needs to sustain itself as a TLP or does it belong as a
subproject under an umbrella project?
Ralph
--------------------------------------------------------------------- To
unsubscribe, e-mail: gene...
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 10:15 AM, Jukka Zitting wrote:
> Hi,
>
> On Wed, Aug 5, 2009 at 6:58 PM, Martijn
> Dashorst<ma...@gmail.com> wrote:
>> AFAIK Nowhere in the policy it is stated that committer activity ==
>> code commits.
>
> True, but that's how I interpret it, for reasons stated earlier in
> this thread.
>
> My -1 in the Pivot vote is just a statement of concern based on my
> interpretation of the exit criteria defined in our policy. If it were
> up to me, I'd ask Pivot to focus a bit more in encouraging others to
> participate also with writing the code before making them a TLP.
>
> You can take that and the other -1 and try to address them before
> calling another vote, or go with the majority result if you think our
> concerns are baseless or already addressed.
>
I don't think your concerns about Pivot are baseless. I think that
stating them in terms of number of commits doesn't really frame your
concern properly.
Here is how I would express my concerns based on the messages I've
read here.
1. We know there are two committers who understand the code very well.
Do the mentors feel the rest of the community has the necessary
understanding and commitment to pick up the slack should something
happen to those two individuals?
2. Does the community participate in the direction of the project? Do
the committers do whatever they want without review or do others
participate?
3. If Pivot were given permission to graduate would the project really
have the momentum it needs to sustain itself as a TLP or does it
belong as a subproject under an umbrella project?
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Thu, Aug 6, 2009 at 5:16 AM, Niclas Hedhman<ni...@hedhman.org> wrote:
> Would it make a difference, if I step up and make the initial PMC Chair, as
> I have over the last couple of weeks started my own commitment to the
> project?
My main concern is over the small number of people writing code for
the project, so I'd actually be more pleased to see you as a committer
(or just a patch contributor) than a non-coding PMC chair.
In fact your recent patch submissions already make the Pivot
contribution report in Jira look better than it used to. Together with
Sandro Martini's recent commits the project is starting to look better
each day.
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Niclas Hedhman <ni...@hedhman.org>.
Jukka,
I don't like to force "majority result" if there is controversy.
Having a couple of -1, and other not voting but practically agreeing with
you and Ant, is not the way to go.
I think they are as ready as they can get, considering the starting point
and the general attitude towards incubated projects. I don't think longer
incubation will improve the chances of sucees, the understanding of
ApacheWay or addressing your concerns (other than artificially)...
Would it make a difference, if I step up and make the initial PMC Chair, as
I have over the last couple of weeks started my own commitment to the
project?
-- Niclas
On Aug 6, 2009 1:31 AM, "Jukka Zitting" <ju...@gmail.com> wrote:
Hi,
On Wed, Aug 5, 2009 at 6:58 PM, Martijn Dashorst<ma...@gmail.com>
wrote: > AFAIK Nowhere...
True, but that's how I interpret it, for reasons stated earlier in this
thread.
My -1 in the Pivot vote is just a statement of concern based on my
interpretation of the exit criteria defined in our policy. If it were
up to me, I'd ask Pivot to focus a bit more in encouraging others to
participate also with writing the code before making them a TLP.
You can take that and the other -1 and try to address them before
calling another vote, or go with the majority result if you think our
concerns are baseless or already addressed.
BR, Jukka Zitting
--------------------------------------------------------------------- To
unsubs...
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Wed, Aug 5, 2009 at 6:58 PM, Martijn
Dashorst<ma...@gmail.com> wrote:
> AFAIK Nowhere in the policy it is stated that committer activity ==
> code commits.
True, but that's how I interpret it, for reasons stated earlier in this thread.
My -1 in the Pivot vote is just a statement of concern based on my
interpretation of the exit criteria defined in our policy. If it were
up to me, I'd ask Pivot to focus a bit more in encouraging others to
participate also with writing the code before making them a TLP.
You can take that and the other -1 and try to address them before
calling another vote, or go with the majority result if you think our
concerns are baseless or already addressed.
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Martijn Dashorst <ma...@gmail.com>.
AFAIK Nowhere in the policy it is stated that committer activity ==
code commits.
Martijn
On Wednesday, August 5, 2009, Jukka Zitting <ju...@gmail.com> wrote:
> Hi,
>
> On Wed, Aug 5, 2009 at 6:02 PM, Ralph Goers<ra...@dslextreme.com> wrote:
>> On Aug 5, 2009, at 8:40 AM, Jukka Zitting wrote:
>>> We are looking for projects with at least three independent
>>> committers, and personally I'm not including inactive committers in
>>> that count.
>>
>> Now you are making hard and fast rules again to predict the future health of
>> the community.
>
> I'm not making up these rules (except as a member of the IPMC). The
> incubation policy explicitly requires "at least 3 legally independent
> committers".
>
> Should we change that to "at least 3 legally independent contributors"?
>
> BR,
>
> Jukka Zitting
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
>
--
Become a Wicket expert, learn from the best: http://wicketinaction.com
Apache Wicket 1.4 increases type safety for web applications
Get it now: http://www.apache.org/dyn/closer.cgi/wicket/1.4.0
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Wed, Aug 5, 2009 at 6:02 PM, Ralph Goers<ra...@dslextreme.com> wrote:
> On Aug 5, 2009, at 8:40 AM, Jukka Zitting wrote:
>> We are looking for projects with at least three independent
>> committers, and personally I'm not including inactive committers in
>> that count.
>
> Now you are making hard and fast rules again to predict the future health of
> the community.
I'm not making up these rules (except as a member of the IPMC). The
incubation policy explicitly requires "at least 3 legally independent
committers".
Should we change that to "at least 3 legally independent contributors"?
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 8:40 AM, Jukka Zitting wrote:
> Hi,
>
> On Wed, Aug 5, 2009 at 5:21 PM, Ralph
> Goers<ra...@dslextreme.com> wrote:
>> You won't find the second part solely from the commit log. I would
>> expect
>> mentors to be monitoring the dev list. If it is filled with "can
>> you fix
>> this?" then there is a problem. But if it also has "why did you do
>> this?",
>> "can we do this?", "I don't understand this commit" type messages
>> then I
>> expect the second concern is adequately addressed.
>
> There's a difference between discussing a change on the list and
> actually committing it to svn. It's the difference between a
> contributor and a committer.
>
> We are looking for projects with at least three independent
> committers, and personally I'm not including inactive committers in
> that count.
>
Now you are making hard and fast rules again to predict the future
health of the community.
I am suggesting that having a community where 2 guys are committing
and several others participate enough that they could take over in the
event 1 should leave is perfectly fine. I'd be far more worried if the
two guys were just doing what they want and not listening to the
others or referring to themselves as project "leads".
Putting up arbitrary hurdles that projects must somehow leap over
isn't the right way to get them out of the nest.
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Wed, Aug 5, 2009 at 5:21 PM, Ralph Goers<ra...@dslextreme.com> wrote:
> You won't find the second part solely from the commit log. I would expect
> mentors to be monitoring the dev list. If it is filled with "can you fix
> this?" then there is a problem. But if it also has "why did you do this?",
> "can we do this?", "I don't understand this commit" type messages then I
> expect the second concern is adequately addressed.
There's a difference between discussing a change on the list and
actually committing it to svn. It's the difference between a
contributor and a committer.
We are looking for projects with at least three independent
committers, and personally I'm not including inactive committers in
that count.
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of Apache Pivot)
Posted by Ralph Goers <ra...@dslextreme.com>.
On Aug 5, 2009, at 2:00 AM, Jukka Zitting wrote:
> Hi,
>
> On Wed, Aug 5, 2009 at 3:39 AM, Ralph
> Goers<ra...@dslextreme.com> wrote:
>> Using these projects as an example is perhaps not the best from a
>> community
>> perspective because Ceki has no intention of running them like Apache
>> projects. But even if he did, by these standards the projects might
>> never
>> make it out of the incubator. Even if those of us who would like
>> them had
>> commit rights I can guarantee that 95% of the commits would still
>> be Ceki's.
>
> I don't see it as a problem if the vast majority of commits comes from
> one person (or company) as long as the community operates normally
> *and* there are others who won't have to start learning how to build
> the codebase and do an "svn commit" if the key developer leaves.
You won't find the second part solely from the commit log. I would
expect mentors to be monitoring the dev list. If it is filled with
"can you fix this?" then there is a problem. But if it also has "why
did you do this?", "can we do this?", "I don't understand this commit"
type messages then I expect the second concern is adequately
addressed. But you can't determine that from raw statistics.
>
> That's why I measure the "three independent committers" criteria by
> looking at the commit log instead of the asf-authorization file. And
> I'm not asking much, just a few code commits in the past few months is
> good enough for me.
>
> That's the criteria that I held Sling against, and that's also what's
> currently keeping UIMA from graduating (and apparently also for over a
> year before I signed up to help them). If the consensus is that this
> is a bit too hard a requirement, then I'll be happy to bring UIMA up
> for graduation in the next few months.
>
I'm suggesting that while that criteria makes sense for someone not
involved in a project, it isn't necessarily the best way to determine
how healthy the community is. If mentors are mentoring then we should
give heavy weight to their recommendation and judgment with respect to
graduation. If a mentor needs these kinds of statistics to tell him/
her whether the project is healthy then I am very concerned.
Ralph
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Thilo Goetz <tw...@gmx.de>.
Jukka Zitting wrote:
> Hi,
>
> On Wed, Aug 5, 2009 at 3:39 AM, Ralph Goers<ra...@dslextreme.com> wrote:
>> Using these projects as an example is perhaps not the best from a community
>> perspective because Ceki has no intention of running them like Apache
>> projects. But even if he did, by these standards the projects might never
>> make it out of the incubator. Even if those of us who would like them had
>> commit rights I can guarantee that 95% of the commits would still be Ceki's.
>
> I don't see it as a problem if the vast majority of commits comes from
> one person (or company) as long as the community operates normally
> *and* there are others who won't have to start learning how to build
> the codebase and do an "svn commit" if the key developer leaves.
>
> That's why I measure the "three independent committers" criteria by
> looking at the commit log instead of the asf-authorization file. And
> I'm not asking much, just a few code commits in the past few months is
> good enough for me.
>
> That's the criteria that I held Sling against, and that's also what's
> currently keeping UIMA from graduating (and apparently also for over a
> year before I signed up to help them). If the consensus is that this
> is a bit too hard a requirement, then I'll be happy to bring UIMA up
> for graduation in the next few months.
Indeed. If this is the way the wind is blowing right now, we're
more than ready for graduation.
--Thilo
>
> BR,
>
> Jukka Zitting
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Wed, Aug 5, 2009 at 3:39 AM, Ralph Goers<ra...@dslextreme.com> wrote:
> Using these projects as an example is perhaps not the best from a community
> perspective because Ceki has no intention of running them like Apache
> projects. But even if he did, by these standards the projects might never
> make it out of the incubator. Even if those of us who would like them had
> commit rights I can guarantee that 95% of the commits would still be Ceki's.
I don't see it as a problem if the vast majority of commits comes from
one person (or company) as long as the community operates normally
*and* there are others who won't have to start learning how to build
the codebase and do an "svn commit" if the key developer leaves.
That's why I measure the "three independent committers" criteria by
looking at the commit log instead of the asf-authorization file. And
I'm not asking much, just a few code commits in the past few months is
good enough for me.
That's the criteria that I held Sling against, and that's also what's
currently keeping UIMA from graduating (and apparently also for over a
year before I signed up to help them). If the consensus is that this
is a bit too hard a requirement, then I'll be happy to bring UIMA up
for graduation in the next few months.
BR,
Jukka Zitting
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org
Re: Diversity as an insurance policy (Was: [VOTE] Graduation of
Apache Pivot)
Posted by Ross Gardler <rg...@apache.org>.
2009/8/5 Ralph Goers <ra...@dslextreme.com>:
>
> On Aug 4, 2009, at 2:10 AM, Jukka Zitting wrote:
>
>> Hi,
>>
>> On Tue, Aug 4, 2009 at 10:10 AM, Bertrand
>> Delacretaz<bd...@apache.org> wrote:
>>>
>>> In such a case, the key point is, do the people who write the code
>>> listen to the community?
>>
>> That's certainly good, but [snip]
...
>> Besides all the other good things diversity brings it's also an
>> insurance policy for the project, and that's what I think the
>> Incubator should be looking for as a graduation criteria.
>>
>
> You are trying to predict the future. Good luck with that.
...
> From my perspective we should be evaluating projects based on whether they
> are building a healthy Apache community where we have sufficient belief that
> the project will be able to sustain itself without further mentoring.
>
+1
a) is there sufficient oversight in the project?
b) will the community actively seek more contributors?
c) does the community know what it knows about the ASF (i.e. will it
ask in unusual circumstances)?
d) have the mentors been active in their mentoring and recommend graduation?
If yes to all these points then we should graduate the project.
Ross
--
Ross Gardler
OSS Watch - supporting open source in education and research
http://www.oss-watch.ac.uk
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org