Jenkins and ASF

[Michael Hüttermann-2]

Hello Andrew, hello all,

according to the gov meeting minutes [1], it is considered to be a
valid strategy to run Jenkins under the ASF umbrella. I've sent a
short tweet about that and got a very overwhelming response from many
different people, including a couple of guys that are activiley
involved in the ASF structure. They offered support to check how to
find a collaboration between Jenkins and ASF. I think hosting Jenkins
as an ASF project would add even more creditability to Jenkins. I
don't think existing licenses prevent such a movement. What do you
think?

Regards
Michael


[1] http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-02-04-23.02.html

[nicolas de loof-2]

Would be a great opportunity for Jenkins, and would solve the  infra / legals issues discussed during meeting.

Apache Extra could be used for plugins that don't use a compatible license.

2011/2/6 Michael Hüttermann <[hidden email]>

Hello Andrew, hello all,

according to the gov meeting minutes [1], it is considered to be a
valid strategy to run Jenkins under the ASF umbrella. I've sent a
short tweet about that and got a very overwhelming response from many
different people, including a couple of guys that are activiley
involved in the ASF structure. They offered support to check how to
find a collaboration between Jenkins and ASF. I think hosting Jenkins
as an ASF project would add even more creditability to Jenkins. I
don't think existing licenses prevent such a movement. What do you
think?

Regards
Michael


[1] http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-02-04-23.02.html

[Nigel Magnay]

Wouldn't that require following ASF release rules, which are a touch archaic?

On Sunday, February 6, 2011, Michael Hüttermann
<[hidden email]> wrote:

> Hello Andrew, hello all,
>
> according to the gov meeting minutes [1], it is considered to be a
> valid strategy to run Jenkins under the ASF umbrella. I've sent a
> short tweet about that and got a very overwhelming response from many
> different people, including a couple of guys that are activiley
> involved in the ASF structure. They offered support to check how to
> find a collaboration between Jenkins and ASF. I think hosting Jenkins
> as an ASF project would add even more creditability to Jenkins. I
> don't think existing licenses prevent such a movement. What do you
> think?
>
> Regards
> Michael
>
>
> [1] http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-02-04-23.02.html

[Thomas Ferris Nicolaisen-2]

I don't have any opinion on this myself, but it did ring a bell over something I read recently: iBatis moved away from Apache to become MyBatis last year.

I'm not aware of all the reasons they did this, but there was at least one legal implication:

When I donated iBATIS to the ASF, it was unfortunately irreversible. This is due to the fact that Apache is a 501(c)(3) nonprofit organization. They cannot turn over what might be considered "assets" to a private entity.

Source: <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%[hidden email]%3E>

Shane of Apache commented this on the fork:

ASF projects are expected to follow a number of procedures roughly known as “The Apache Way”. We believe that software projects with a diverse community; that use a consensus process to make decisions; and that do all their work in the open results in the best overall quality and longest lasting projects. While the ASF expects it’s projects to follow the Apache Way, we certainly understand that not everyone believes that our way of software development is the right one for everyone or for all projects.

Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>

I don't know what's been going on behind the curtains there, but it's certainly important to make sure you're aware of Apache's expectations and constraints before you make the move (well, stating the obivous there).

[pelegri]

I have mixed feelings about "the Apache way".  It can work very well,
but it can also lead to internal fights, paralysis, conflicting
agendas.  In particular Sun - where I was - was a founding member of
Tomcat but eventually felt it had to go its own way because of some of
these problems.  And when we designed the GlassFish governance, we
chose a "(Corporate) benevolent dictatorship" model - which has its
own problems.

But, in any case, I don't see how Jenkins can be an Apache project.
Apache projects require IP to be contributed, and the bulk of the pre-
Jenkins IP is (c) Oracle, so we can't contribute it.

There is Apache-Extras at Google, but these are not really ASF
projects.  As far as I understand the arrangement, there is no formal
relationshp between any project at Apache-Extras and ASF and it would
not really address any of our issues.

http://code.google.com/a/apache-extras.org/p/apache-extras/wiki/FAQ

  Eduard/o

On Feb 6, 11:24 am, Thomas Ferris Nicolaisen <[hidden email]> wrote:

> I don't have any opinion on this myself, but it did ring a bell over
> something I read recently: iBatis moved away from Apache to become MyBatis
> last year.
>
> I'm not aware of all the reasons they did this, but there was at least one
> legal implication:
>
> When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> cannot turn over what might be considered "assets" to a private entity.
>
> Source:
> <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>
>
> Shane of Apache commented this on the fork:
>
> ASF projects are expected to follow a number of procedures roughly known as
> “The Apache Way”. We believe that software projects with a diverse
> community; that use a consensus process to make decisions; and that do all
> their work in the open results in the best overall quality and longest
> lasting projects. While the ASF expects it’s projects to follow the Apache
> Way, we certainly understand that not everyone believes that our way of
> software development is the right one for everyone or for all projects.
>
> Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> I don't know what's been going on behind the curtains there, but it's
> certainly important to make sure you're aware of Apache's expectations and
> constraints before you make the move (well, stating the obivous there).

[nicolas de loof-2]

Regarding the "Apache way", I've noticed from Apache Maven & Apache commons where I contributed that only the PMC is responsible for project governance (nut is encouraged to get other contributors/committers involved), based on a consensus (the famous +1 / -1 votes). PMC can be limited in size, so it may be constituted on historical leadership on Hudson project and be similar to a 3-heads benevolent dictator model.

The main issue is not the "Apache way", that may fail only when the community has other governance/leadership issues, but (as you notice) the strict Apache IP rules. I don't know well the incubator requirements at apache to get a project donated to the foundation, but this may fail with Oracle copyright on code ( http://incubator.apache.org/ip-clearance/index.html ) even if I thing the MIT license is safe enough in such case. Maybe we could ask [hidden email]

Nicolas

2011/2/6 pelegri <[hidden email]>

I have mixed feelings about "the Apache way".  It can work very well,
but it can also lead to internal fights, paralysis, conflicting
agendas.  In particular Sun - where I was - was a founding member of
Tomcat but eventually felt it had to go its own way because of some of
these problems.  And when we designed the GlassFish governance, we
chose a "(Corporate) benevolent dictatorship" model - which has its
own problems.

But, in any case, I don't see how Jenkins can be an Apache project.
Apache projects require IP to be contributed, and the bulk of the pre-
Jenkins IP is (c) Oracle, so we can't contribute it.

There is Apache-Extras at Google, but these are not really ASF
projects.  As far as I understand the arrangement, there is no formal
relationshp between any project at Apache-Extras and ASF and it would
not really address any of our issues.

http://code.google.com/a/apache-extras.org/p/apache-extras/wiki/FAQ

 Eduard/o

On Feb 6, 11:24 am, Thomas Ferris Nicolaisen <[hidden email]> wrote:
> I don't have any opinion on this myself, but it did ring a bell over
> something I read recently: iBatis moved away from Apache to become MyBatis
> last year.
>
> I'm not aware of all the reasons they did this, but there was at least one
> legal implication:
>
> When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> cannot turn over what might be considered "assets" to a private entity.
>
> Source:

> <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>

>
> Shane of Apache commented this on the fork:
>
> ASF projects are expected to follow a number of procedures roughly known as
> “The Apache Way”. We believe that software projects with a diverse
> community; that use a consensus process to make decisions; and that do all
> their work in the open results in the best overall quality and longest
> lasting projects. While the ASF expects it’s projects to follow the Apache
> Way, we certainly understand that not everyone believes that our way of
> software development is the right one for everyone or for all projects.
>
> Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> I don't know what's been going on behind the curtains there, but it's
> certainly important to make sure you're aware of Apache's expectations and
> constraints before you make the move (well, stating the obivous there).

[Andrew Bayer]

FWIW, I've taken advantage of my employer also employing Doug Cutting, ASF chair, and have emailed him to see what if any options are available for a Jenkins/Apache relationship. I'll let you all know how that goes.

A.

On Sun, Feb 6, 2011 at 12:31 PM, nicolas de loof <[hidden email]> wrote:

Regarding the "Apache way", I've noticed from Apache Maven & Apache commons where I contributed that only the PMC is responsible for project governance (nut is encouraged to get other contributors/committers involved), based on a consensus (the famous +1 / -1 votes). PMC can be limited in size, so it may be constituted on historical leadership on Hudson project and be similar to a 3-heads benevolent dictator model.

The main issue is not the "Apache way", that may fail only when the community has other governance/leadership issues, but (as you notice) the strict Apache IP rules. I don't know well the incubator requirements at apache to get a project donated to the foundation, but this may fail with Oracle copyright on code ( http://incubator.apache.org/ip-clearance/index.html ) even if I thing the MIT license is safe enough in such case. Maybe we could ask [hidden email]

Nicolas

2011/2/6 pelegri <[hidden email]>

I have mixed feelings about "the Apache way".  It can work very well,
but it can also lead to internal fights, paralysis, conflicting
agendas.  In particular Sun - where I was - was a founding member of
Tomcat but eventually felt it had to go its own way because of some of
these problems.  And when we designed the GlassFish governance, we
chose a "(Corporate) benevolent dictatorship" model - which has its
own problems.

But, in any case, I don't see how Jenkins can be an Apache project.
Apache projects require IP to be contributed, and the bulk of the pre-
Jenkins IP is (c) Oracle, so we can't contribute it.

There is Apache-Extras at Google, but these are not really ASF
projects.  As far as I understand the arrangement, there is no formal
relationshp between any project at Apache-Extras and ASF and it would
not really address any of our issues.

http://code.google.com/a/apache-extras.org/p/apache-extras/wiki/FAQ

 Eduard/o

On Feb 6, 11:24 am, Thomas Ferris Nicolaisen <[hidden email]> wrote:
> I don't have any opinion on this myself, but it did ring a bell over
> something I read recently: iBatis moved away from Apache to become MyBatis
> last year.
>
> I'm not aware of all the reasons they did this, but there was at least one
> legal implication:
>
> When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> cannot turn over what might be considered "assets" to a private entity.
>
> Source:

> <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>

>
> Shane of Apache commented this on the fork:
>
> ASF projects are expected to follow a number of procedures roughly known as
> “The Apache Way”. We believe that software projects with a diverse
> community; that use a consensus process to make decisions; and that do all
> their work in the open results in the best overall quality and longest
> lasting projects. While the ASF expects it’s projects to follow the Apache
> Way, we certainly understand that not everyone believes that our way of
> software development is the right one for everyone or for all projects.
>
> Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> I don't know what's been going on behind the curtains there, but it's
> certainly important to make sure you're aware of Apache's expectations and
> constraints before you make the move (well, stating the obivous there).

[rgardler]

Just an intro of myself...

I'm not a developer on Jenkins (or Hudson) and have never even read a
mail on your lists prior to this one. Therefore I have no opinion
about whether you should/should not enter the ASF.

I've joined this thread in order to help you understand what it means
from the ASF side of things. I'm active in the incubator and in a few
other things at the ASF.

Please note I'm not subscribed to this list with one of my primary
emails so I may be slow in responding. If you have something specific
you want me to respond to copy just ping me at rgardler....apache.org
and I'll be sure to come back and review the thread.

I'll respond to a couple of points elsewhere in this thread now.

Ross

On Feb 6, 4:40 pm, Michael Hüttermann
<[hidden email]> wrote:

> Hello Andrew, hello all,
>
> according to the gov meeting minutes [1], it is considered to be a
> valid strategy to run Jenkins under the ASF umbrella. I've sent a
> short tweet about that and got a very overwhelming response from many
> different people, including a couple of guys that are activiley
> involved in the ASF structure. They offered support to check how to
> find a collaboration between Jenkins and ASF. I think hosting Jenkins
> as an ASF project would add even more creditability to Jenkins. I
> don't think existing licenses prevent such a movement. What do you
> think?
>
> Regards
> Michael
>
> [1]http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-02-04-23.02....

[rgardler]

On Feb 6, 7:20 pm, Nigel Magnay <[hidden email]> wrote:
> Wouldn't that require following ASF release rules, which are a touch archaic?

It depends how you define "archaic". If you mean sufficiently robust
to ensure that big corporation lawyers will allow Apache code to be
used in their products then yes, then no they are not "archaic" they
are "robust".

If you mean  some people don't like them, then yes you are right.

Either way, all ASF projects must adhere to our release rules, it is
one of the very few things that are non-negotiable in the ASF. There
is more on this and a few other fixed rules at
https://blogs.apache.org/comdev/entry/what_makes_apache_projects_different

Ross

>
> On Sunday, February 6, 2011, Michael Hüttermann
>
> <[hidden email]> wrote:
> > Hello Andrew, hello all,
>
> > according to the gov meeting minutes [1], it is considered to be a
> > valid strategy to run Jenkins under the ASF umbrella. I've sent a
> > short tweet about that and got a very overwhelming response from many
> > different people, including a couple of guys that are activiley
> > involved in the ASF structure. They offered support to check how to
> > find a collaboration between Jenkins and ASF. I think hosting Jenkins
> > as an ASF project would add even more creditability to Jenkins. I
> > don't think existing licenses prevent such a movement. What do you
> > think?
>
> > Regards
> > Michael
>
> > [1]http://meetings.jenkins-ci.org/jenkins/2011/jenkins.2011-02-04-23.02....

[rgardler]

On Feb 6, 7:24 pm, Thomas Ferris Nicolaisen <[hidden email]> wrote:

> I don't have any opinion on this myself, but it did ring a bell over
> something I read recently: iBatis moved away from Apache to become MyBatis
> last year.
>
> I'm not aware of all the reasons they did this, but there was at least one
> legal implication:
>
> When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> cannot turn over what might be considered "assets" to a private entity.

To be clear, this applies to the Trademark, not to the code. The code
is under an Apache License which is sufficiently permissive that you
can do whatever you want with it outside of the foundation (although
you still have to abide by the Apache Licence itself).

> Source:
> <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>
>
> Shane of Apache commented this on the fork:
>
> ASF projects are expected to follow a number of procedures roughly known as
> “The Apache Way”. We believe that software projects with a diverse
> community; that use a consensus process to make decisions; and that do all
> their work in the open results in the best overall quality and longest
> lasting projects. While the ASF expects it’s projects to follow the Apache
> Way, we certainly understand that not everyone believes that our way of
> software development is the right one for everyone or for all projects.
>
> Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> I don't know what's been going on behind the curtains there, but it's
> certainly important to make sure you're aware of Apache's expectations and
> constraints before you make the move (well, stating the obivous there).

There's nothing behind the curtains. The iBatis project team didn't
like the way the ASF ran things and decided to go their own way.

I do agree that it is important that you understand what the ASF is
about. There are few rules that we insist upon, but those few rules
are not to everyones liking.

Ross

[rgardler]

On Feb 6, 7:53 pm, pelegri <[hidden email]> wrote:
> I have mixed feelings about "the Apache way".  It can work very well,
> but it can also lead to internal fights, paralysis, conflicting
> agendas.  In particular Sun - where I was - was a founding member of
> Tomcat but eventually felt it had to go its own way because of some of
> these problems.  And when we designed the GlassFish governance, we
> chose a "(Corporate) benevolent dictatorship" model - which has its
> own problems.

A well run community does not have "internal fights, paralysis
conflicting agendas". Internal fights are a symptom of a lack of
respect or understanding between opposing views. This can occur in any
community that is not well run. It is not a symptom of the Apache Way.
Paralysis is extremely rare in the ASF due to the way we make
decisions. It can happen in extreme cases, but in those cases I would
argue the same as for "internal fights". Conflicting Agendas are a
fact of life, and again not caused by the Apache Way. The question is
does the Apache Way provide a means to resolve those conflicts, again
I refer to my first two points.

Sometimes conflicts are not resolvable, at this point a fork is
created - that's what open source is all about (err.... hello Hudson,
oh sorry Jenkins).

Like most well run open source projects forks of ASF code are unusual,
but they do occasionally occur.

As Eduardo points out no governance model is perfeect. There are some
good documents (well, I wrote them so I think they are good) on
governance models at:

http://www.oss-watch.ac.uk/resources/governanceModels.xml
http://www.oss-watch.ac.uk/resources/benevolentdictatorgovernancemodel.xml
http://www.oss-watch.ac.uk/resources/meritocraticGovernanceModel.xml

Note that a well run benevolent dictatorship is actually very similar
to a well run meritocratic model. Some of the processes are different
(in particular conflict resolution) but community engagement and
collaboration models are almost identical.

> But, in any case, I don't see how Jenkins can be an Apache project.
> Apache projects require IP to be contributed, and the bulk of the pre-
> Jenkins IP is (c) Oracle, so we can't contribute it.

No, this is not correct. The ASF does not require copyrights. What we
do require though is a copyright license. If Oracle are not prepared
to give us that licence via our standard Contributor Licence Agreement
we would have to work with out legal team to figure out how this could
be done. As I understand it the code is MIT licensed, this is
compatible with the Apache Licence.

However, this does not mean that we would be able to move the code
over. I've asked out legal list for their thoughts:
http://markmail.org/thread/xpsgspwozy2aiotv

> There is Apache-Extras at Google, but these are not really ASF
> projects.  As far as I understand the arrangement, there is no formal
> relationshp between any project at Apache-Extras and ASF and it would
> not really address any of our issues.

This is correct. Apache-Extras is for projects that are related to the
ASF, but not owned by the ASF. It's really intended for things like
plugins for Apache projects that cannot be licenced under the Apache
licence. I don't think you would gain any significant benefit from
being there over being in Google Code itself.

Ross

>
> http://code.google.com/a/apache-extras.org/p/apache-extras/wiki/FAQ
>
>   Eduard/o
>
> On Feb 6, 11:24 am, Thomas Ferris Nicolaisen <[hidden email]> wrote:
>
> > I don't have any opinion on this myself, but it did ring a bell over
> > something I read recently: iBatis moved away from Apache to become MyBatis
> > last year.
>
> > I'm not aware of all the reasons they did this, but there was at least one
> > legal implication:
>
> > When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> > due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> > cannot turn over what might be considered "assets" to a private entity.
>
> > Source:
> > <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>
>
> > Shane of Apache commented this on the fork:
>
> > ASF projects are expected to follow a number of procedures roughly known as
> > “The Apache Way”. We believe that software projects with a diverse
> > community; that use a consensus process to make decisions; and that do all
> > their work in the open results in the best overall quality and longest
> > lasting projects. While the ASF expects it’s projects to follow the Apache
> > Way, we certainly understand that not everyone believes that our way of
> > software development is the right one for everyone or for all projects.
>
> > Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> > I don't know what's been going on behind the curtains there, but it's
> > certainly important to make sure you're aware of Apache's expectations and
> > constraints before you make the move (well, stating the obivous there).

[pelegri]

On the trademark issue, my understanding is that the only  releases
that can use the ASF-branded name are exactly those from ASF.

That is why, for example, all the vendors that distribute tomcat-based
distributions use a different name for their distributions.

  eduard/o

On Feb 6, 3:33 pm, rgardler <[hidden email]> wrote:

> On Feb 6, 7:24 pm, Thomas Ferris Nicolaisen <[hidden email]> wrote:
>
> > I don't have any opinion on this myself, but it did ring a bell over
> > something I read recently: iBatis moved away from Apache to become MyBatis
> > last year.
>
> > I'm not aware of all the reasons they did this, but there was at least one
> > legal implication:
>
> > When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> > due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> > cannot turn over what might be considered "assets" to a private entity.
>
> To be clear, this applies to the Trademark, not to the code. The code
> is under an Apache License which is sufficiently permissive that you
> can do whatever you want with it outside of the foundation (although
> you still have to abide by the Apache Licence itself).
>
>
>
> > Source:
> > <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>
>
> > Shane of Apache commented this on the fork:
>
> > ASF projects are expected to follow a number of procedures roughly known as
> > “The Apache Way”. We believe that software projects with a diverse
> > community; that use a consensus process to make decisions; and that do all
> > their work in the open results in the best overall quality and longest
> > lasting projects. While the ASF expects it’s projects to follow the Apache
> > Way, we certainly understand that not everyone believes that our way of
> > software development is the right one for everyone or for all projects.
>
> > Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> > I don't know what's been going on behind the curtains there, but it's
> > certainly important to make sure you're aware of Apache's expectations and
> > constraints before you make the move (well, stating the obivous there).
>
> There's nothing behind the curtains. The iBatis project team didn't
> like the way the ASF ran things and decided to go their own way.
>
> I do agree that it is important that you understand what the ASF is
> about. There are few rules that we insist upon, but those few rules
> are not to everyones liking.
>
> Ross

[rgardler]

On Feb 6, 8:31 pm, nicolas de loof <[hidden email]> wrote:
> Regarding the "Apache way", I've noticed from Apache Maven & Apache commons
> where I contributed that only the PMC is responsible for project governance
> (nut is encouraged to get other contributors/committers involved), based on
> a consensus (the famous +1 / -1 votes). PMC can be limited in size, so it
> may be constituted on historical leadership on Hudson project and be similar
> to a 3-heads benevolent dictator model.

No we would not allow that. All Apache projects are meritocracies and
must be run as one. Having a hard fix on the number of people able to
become a part of the governance of the project is not meritocratic.
This is another of the invariant rules of the ASF.

Having said that, each project is free to decide where to place the
bar for membership of the PMC. Some projects give it away quickly (a
couple of decent patches for example), others require a significant
amount of time and effort. As long as the rules the project defines
are meritocratic (i.e. there is a way to become a part of the PMC)
then they will be accepted by the board. Note, however, that the
culture of the ASF leans more towards low barriers than high barriers.
Experience has shown us that lower barriers are more effective in
creating viable ASF style communities - of course, I don't mean to
imply that ASF style communities are the only succesful ones.

Projects are also free to choose whether they want to make the
distinction between committer and PMC member. Historically there was
always a split. Over the years many projects (though not all) have
decided that committership == PMC. There are still projects that have
a clear separation between committer and PMC member.

A further area of flexibility is in who makes the decisions. In a
healthy ASF style community decisions are made through a consensus
which is reached through code and (where necessary) discussion.
Conflicts are resolved with a vote. How those votes are counted is up
to the PMC to define (with a few exceptions that have legal
implications).

> The main issue is not the "Apache way", that may fail only when the
> community has other governance/leadership issues, but (as you notice) the
> strict Apache IP rules. I don't know well the incubator requirements at
> apache to get a project donated to the foundation, but this may fail with
> Oracle copyright on code (http://incubator.apache.org/ip-clearance/index.html) even if I thing the
> MIT license is safe enough in such case. Maybe we could ask
> [hidden email]

I already asked for you. See http://markmail.org/thread/xpsgspwozy2aiotv

Ross

[rgardler]

On Feb 6, 11:56 pm, pelegri <[hidden email]> wrote:
> On the trademark issue, my understanding is that the only  releases
> that can use the ASF-branded name are exactly those from ASF.
>
> That is why, for example, all the vendors that distribute tomcat-based
> distributions use a different name for their distributions.

Correct. More at http://www.apache.org/foundation/marks/

Ross

>
>   eduard/o
>
> On Feb 6, 3:33 pm, rgardler <[hidden email]> wrote:
>
> > On Feb 6, 7:24 pm, Thomas Ferris Nicolaisen <[hidden email]> wrote:
>
> > > I don't have any opinion on this myself, but it did ring a bell over
> > > something I read recently: iBatis moved away from Apache to become MyBatis
> > > last year.
>
> > > I'm not aware of all the reasons they did this, but there was at least one
> > > legal implication:
>
> > > When I donated iBATIS to the ASF, it was unfortunately irreversible. This is
> > > due to the fact that Apache is a 501(c)(3) nonprofit organization. They
> > > cannot turn over what might be considered "assets" to a private entity.
>
> > To be clear, this applies to the Trademark, not to the code. The code
> > is under an Apache License which is sufficiently permissive that you
> > can do whatever you want with it outside of the foundation (although
> > you still have to abide by the Apache Licence itself).
>
> > > Source:
> > > <http://mail-archives.apache.org/mod_mbox/ibatis-user-java/201005.mbox/%3CAANLkTilKGrPRn48gAhCWkdrLWQFGkXOkjbjHYVjsW...@...%3E>
>
> > > Shane of Apache commented this on the fork:
>
> > > ASF projects are expected to follow a number of procedures roughly known as
> > > “The Apache Way”. We believe that software projects with a diverse
> > > community; that use a consensus process to make decisions; and that do all
> > > their work in the open results in the best overall quality and longest
> > > lasting projects. While the ASF expects it’s projects to follow the Apache
> > > Way, we certainly understand that not everyone believes that our way of
> > > software development is the right one for everyone or for all projects.
>
> > > Source: <http://communityovercode.com/2010/06/mybatis-forks-apache-ibatis/>
>
> > > I don't know what's been going on behind the curtains there, but it's
> > > certainly important to make sure you're aware of Apache's expectations and
> > > constraints before you make the move (well, stating the obivous there).
>
> > There's nothing behind the curtains. The iBatis project team didn't
> > like the way the ASF ran things and decided to go their own way.
>
> > I do agree that it is important that you understand what the ASF is
> > about. There are few rules that we insist upon, but those few rules
> > are not to everyones liking.
>
> > Ross

[pelegri]

rgardler wrote

As Eduardo points out no governance model is perfeect. There are some
good documents (well, I wrote them so I think they are good) on
governance models at:

http://www.oss-watch.ac.uk/resources/governanceModels.xml
http://www.oss-watch.ac.uk/resources/benevolentdictatorgovernancemodel.xml
http://www.oss-watch.ac.uk/resources/meritocraticGovernanceModel.xml

Thanks for the pointers, Ross.

rgardler wrote

> But, in any case, I don't see how Jenkins can be an Apache project.
> Apache projects require IP to be contributed, and the bulk of the pre-
> Jenkins IP is (c) Oracle, so we can't contribute it.

No, this is not correct. The ASF does not require copyrights. What we
do require though is a copyright license. If Oracle are not prepared
to give us that licence via our standard Contributor Licence Agreement
we would have to work with out legal team to figure out how this could
be done. As I understand it the code is MIT licensed, this is
compatible with the Apache Licence.

However, this does not mean that we would be able to move the code
over. I've asked out legal list for their thoughts:
http://markmail.org/thread/xpsgspwozy2aiotv

Thanks!

[Andrew Bayer]

For what it's worth, my only real concern is how to mesh the Apache release policies with Jenkins' rapid iteration of releases - that said, I'm not entirely sure it'd be a bad thing for the release cycle to slow down a bit (not much, but not necessarily a release a week, every week).

A.

On Sun, Feb 6, 2011 at 4:37 PM, pelegri <[hidden email]> wrote:

rgardler wrote:
>
> As Eduardo points out no governance model is perfeect. There are some
> good documents (well, I wrote them so I think they are good) on
> governance models at:
>
> http://www.oss-watch.ac.uk/resources/governanceModels.xml
> http://www.oss-watch.ac.uk/resources/benevolentdictatorgovernancemodel.xml
> http://www.oss-watch.ac.uk/resources/meritocraticGovernanceModel.xml
>

Thanks for the pointers, Ross.

rgardler wrote:
>
>> But, in any case, I don't see how Jenkins can be an Apache project.
>> Apache projects require IP to be contributed, and the bulk of the pre-
>> Jenkins IP is (c) Oracle, so we can't contribute it.
>
> No, this is not correct. The ASF does not require copyrights. What we
> do require though is a copyright license. If Oracle are not prepared
> to give us that licence via our standard Contributor Licence Agreement
> we would have to work with out legal team to figure out how this could
> be done. As I understand it the code is MIT licensed, this is
> compatible with the Apache Licence.
>
> However, this does not mean that we would be able to move the code
> over. I've asked out legal list for their thoughts:
> http://markmail.org/thread/xpsgspwozy2aiotv
>

Thanks!

--
View this message in context: http://jenkins.361315.n4.nabble.com/Jenkins-and-ASF-tp3262928p3263419.html
Sent from the Jenkins dev mailing list archive at Nabble.com.

[Richard Bywater]

Couldn't the weekly releases be a bit like Release Candidate type releases (well thats probably not the right name but hopefully it makes sense) and then there's a official release at a more irregular interval - that could go towards the issue of a "Stable" release too if done correctly?

Richard.

On Mon, Feb 7, 2011 at 2:50 PM, Andrew Bayer <[hidden email]> wrote:

For what it's worth, my only real concern is how to mesh the Apache release policies with Jenkins' rapid iteration of releases - that said, I'm not entirely sure it'd be a bad thing for the release cycle to slow down a bit (not much, but not necessarily a release a week, every week).

A.

On Sun, Feb 6, 2011 at 4:37 PM, pelegri <[hidden email]> wrote:

rgardler wrote:
>
> As Eduardo points out no governance model is perfeect. There are some
> good documents (well, I wrote them so I think they are good) on
> governance models at:
>
> http://www.oss-watch.ac.uk/resources/governanceModels.xml
> http://www.oss-watch.ac.uk/resources/benevolentdictatorgovernancemodel.xml
> http://www.oss-watch.ac.uk/resources/meritocraticGovernanceModel.xml
>

Thanks for the pointers, Ross.

rgardler wrote:
>
>> But, in any case, I don't see how Jenkins can be an Apache project.
>> Apache projects require IP to be contributed, and the bulk of the pre-
>> Jenkins IP is (c) Oracle, so we can't contribute it.
>
> No, this is not correct. The ASF does not require copyrights. What we
> do require though is a copyright license. If Oracle are not prepared
> to give us that licence via our standard Contributor Licence Agreement
> we would have to work with out legal team to figure out how this could
> be done. As I understand it the code is MIT licensed, this is
> compatible with the Apache Licence.
>
> However, this does not mean that we would be able to move the code
> over. I've asked out legal list for their thoughts:
> http://markmail.org/thread/xpsgspwozy2aiotv
>

Thanks!

--
View this message in context: http://jenkins.361315.n4.nabble.com/Jenkins-and-ASF-tp3262928p3263419.html
Sent from the Jenkins dev mailing list archive at Nabble.com.

[pelegri]

rgardler wrote

On Feb 6, 8:31 pm, nicolas de loof <[e-mail]> wrote:
> Regarding the "Apache way", I've noticed from Apache Maven & Apache commons
> where I contributed that only the PMC is responsible for project governance
> (nut is encouraged to get other contributors/committers involved), based on
> a consensus (the famous +1 / -1 votes). PMC can be limited in size, so it
> may be constituted on historical leadership on Hudson project and be similar
> to a 3-heads benevolent dictator model.

No we would not allow that. All Apache projects are meritocracies and
must be run as one. Having a hard fix on the number of people able to
become a part of the governance of the project is not meritocratic.
This is another of the invariant rules of the ASF.

How flexibly is ASF's definition of "merit"?

I am totally fine with an egalitarian approach to votes, etc.  My main concern with the ASF model - and I might be misrepresenting the model - is that ASF meritocracy seems to mean just "technical expertise" and I think it is important to consider how well the team works together.

[Steve M. Robbins]

On Sun, Feb 06, 2011 at 05:50:08PM -0800, Andrew Bayer wrote:
> For what it's worth, my only real concern is how to mesh the Apache release
> policies with Jenkins' rapid iteration of releases - that said, I'm not
> entirely sure it'd be a bad thing for the release cycle to slow down a bit
> (not much, but not necessarily a release a week, every week).

Well, reading from the Apache Releases FAQ [1], it is completely
permissible to offer "Test Packages" without a formal approval from
the PMC.  One could imagine the weekly build be such a Test Package.

On the other hand: as a user, I'd certainly welcome a slower release
cycle coupled with automated testing and some QA.  This should give me
some more assurance that an upgrade would be successful.

[1] http://www.apache.org/dev/release.html

Cheers,
-Steve

[Jeff Jensen-2]

On Sun, Feb 6, 2011 at 8:11 PM, pelegri <[hidden email]> wrote:

>
>
> rgardler wrote:
>>
>> On Feb 6, 8:31 pm, nicolas de loof <[hidden email]> wrote:
>>> Regarding the "Apache way", I've noticed from Apache Maven & Apache
>>> commons
>>> where I contributed that only the PMC is responsible for project
>>> governance
>>> (nut is encouraged to get other contributors/committers involved), based
>>> on
>>> a consensus (the famous +1 / -1 votes). PMC can be limited in size, so it
>>> may be constituted on historical leadership on Hudson project and be
>>> similar
>>> to a 3-heads benevolent dictator model.
>>
>> No we would not allow that. All Apache projects are meritocracies and
>> must be run as one. Having a hard fix on the number of people able to
>> become a part of the governance of the project is not meritocratic.
>> This is another of the invariant rules of the ASF.
>>
>
> How flexibly is ASF's definition of "merit"?

Very.  An existing committer "sponsors" you, and the committers vote
+1, 0, -1 to grant you committer status.  Any -1 votes mean you don't
receive it.
You don't typically get nominated/sponsored without first proving your
worth one way or another.


> I am totally fine with an egalitarian approach to votes, etc.  My main
> concern with the ASF model - and I might be misrepresenting the model - is
> that ASF meritocracy seems to mean just "technical expertise" and I think it
> is important to consider how well the team works together.

It is not just technical expertise.


I think the biggest issue with moving to Apache is the requirement for
votes to release with a voting period of 72 hours.  Continuing a
weekly-ish release could be overcome by staging a release and vote on
every Tues for a Fri release.

The rigor and structure provided by Apache could be very beneficial to
Jenkins.  It would also be painful with some of its process
requirements.

Regardless of Apache or not, the key need is the proven release
quality - test coverage, process, infrastructure.