Hi Geert,

> On Oct 11, 2016, at 12:10 AM, Geertjan Wielenga 
> <[email protected]> wrote:
> 
> Here's what it looks like:
> http://www.apache.org/licenses/software-grant-template.pdf

Please do not use this license. Use the CCLA instead.

The CCLA clarifies the status of all Oracle employees with regard to their 
contributions to the code.

http://www.apache.org/licenses/cla-corporate.txt

Thanks,

Craig

> 
> It's in the process of being signed right now, it's being worked on right
> now, might take a week or so the way it looks now.
> 
> The question remains -- and can someone answer it: once the grant has been
> signed and handed over to Apache, what happens if for some reason the
> process fails, must Apache then sign a document to grant the code back to
> Oracle?
> 
> Gj
> 
> 
> On Tue, Oct 11, 2016 at 8:34 AM, Emilian Bold <[email protected]>
> wrote:
> 
>> I didn't mean just an empty git repo, I meant the canonical repository from
>> which daily builds and releases are made.
>> 
>> I believe with this proposal Oracle has agreed to the following:
>> 
>> 1. Changing the project license to the Apache license
>> 2. Contributing further changes under the Apache license
>> 3. Following the Apache governance model and
>> 3. Granting code ownership to the Apache Software Foundation.
>> 
>> I don't know how a software grant document looks like but I assume there
>> are articles about 'unwinding'. Oracle legal should talk to Apache legal
>> and clear this out.
>> 
>> It seems to me though that without the code grant incubation hasn't really
>> started. I mean, incubation is not about due diligence or legal discovery.
>> 
>> Still, there is nothing stopping Oracle from following 1, 2 and 3. They
>> could change the license to the Apache license this very week.
>> 
>> 
>> 
>> --emi
>> 
>> On Tue, Oct 11, 2016 at 5:46 AM, Geertjan Wielenga <
>> [email protected]> wrote:
>> 
>>> The point is this -- during incubation, we're going to be working on
>>> establishing whether Apache NetBeans can exist or not, from many
>> different
>>> points of view. And, even though we don't believe the process will fail,
>> it
>>> would be a problem if Oracle has granted the code to Apache only to find
>>> that for some reason Apache NetBeans will not be able to leave the
>>> incubator. Let's say, for example, there's a licensing problem that
>> cannot
>>> be fixed. If the software has already been granted, it would then need to
>>> be 'ungranted' at that stage. That's my concern and why I think the code
>>> should only be granted formally, i.e., via the formal SGA document, when
>> we
>>> know for sure that incubation will succeed.
>>> 
>>> That means that we can work on setting up the Git repo immediately and,
>>> once we know what we want to move there, we move the source code there.
>>> Then we start the process of 'scrubbing the code', i.e., checking its
>>> licenses and noting any problems and seeking their solutions. Not sure
>> how
>>> long this will take, but maybe not too long, a month or so, just a
>>> guesstimate. Once we have worked through the licensing, and we know for
>>> sure incubation will succeed, we can get the SGA, if we know for sure
>> there
>>> will be no blockers. We did a preliminary investigation of this prior to
>>> putting the proposal together, but at this point we'll have done a
>> thorough
>>> analysis.
>>> 
>>> Then, once we have the SGA, those who have signed the ICLAs can begin
>>> working on committing code agreed upon by the project in terms of a
>>> commonly drawn up roadmap. So, it's not a question of waiting until next
>>> year sometime to start committing, just a question of waiting until we
>> know
>>> for 100% sure that the process will not have to be unwound before
>> actually
>>> having the code granted from Oracle.
>>> 
>>> Does the above make sense?
>>> 
>>> Gj
>>> 
>>> 
>>> On Tue, Oct 11, 2016 at 1:29 AM, Emilian Bold <[email protected]>
>>> wrote:
>>> 
>>>> Migrating the repository over to git and the code grant should happen
>> in
>>>> 2016.
>>>> 
>>>> We have some momentum here but if I have to wait until Summer 2017 to
>>>> commit using my @apache ID I signed the iCLA 6 months too soon.
>>>> 
>>>> Also, it's a premature optimization to change too much the code
>>> repository.
>>>> It seems like a juicy engineering task to split it up, filter it,
>>> whatever.
>>>> But it is pointless.
>>>> 
>>>> What's essential first is for work to be possible and to start on the
>> git
>>>> repo. We could have another goal during the incubation or even after
>>>> incubation to split the repository.
>>>> 
>>>> I don't think the unwinding should be your main concern. Code changes
>>> will
>>>> have to be done regardless of who owns the IP.
>>>> 
>>>> As an alternative to this Oracle concern, you could require
>> contributors
>>> to
>>>> have both an iCLA and an OCA, although perhaps the Apache iCLA might be
>>>> sufficient. Apache Legal might intervene and explain things here...
>>>> 
>>>> An incubating project must do a major release during incubation. I
>>> believe
>>>> that release will have be the Java 9 release.
>>>> 
>>>> 
>>>> 
>>>> --emi
>>>> 
>>>> On Tue, Oct 11, 2016 at 12:18 AM, Geertjan Wielenga <
>>>> [email protected]> wrote:
>>>> 
>>>>> Hi all,
>>>>> 
>>>>> An overview of the sequence as far as I understand it. Consider it a
>>>> basic
>>>>> starting point for discussion.
>>>>> 
>>>>> Let's start by assuming we want there to be a NetBeans 9 to be
>> released
>>>> out
>>>>> of Apache, and as a top level project, i.e., outside the incubator,
>> in
>>>> line
>>>>> with the release of Java 9.
>>>>> 
>>>>> That puts us in the middle of next year somewhere.
>>>>> 
>>>>> The most important aspect that needs to be worked through before then
>>> is
>>>>> the IP, license hygiene, etc. Before we get to the point where we're
>>>>> working on that, we need to actually have one or more Mercurial repos
>>>> that
>>>>> we know we want to move. Right now, the NetBeans 9 branch is being
>>> moved
>>>>> into trunk, once that's done we need to consider whether we should
>> take
>>>> the
>>>>> NetBeans trunk as our starting point -- and determine other brances
>>> we'll
>>>>> need.
>>>>> 
>>>>> We'll then need to work through the IP issues, i.e., work through the
>>>>> incompatible licenses and work out solutions for those. Some features
>>>> might
>>>>> be dropped, others can be installed via plugins, either separately or
>>>>> during installation.
>>>>> 
>>>>> At the point where we've worked through those licensing issues and
>> are
>>>> at a
>>>>> stage where we either have temporary exceptions for truly problematic
>>>>> areas, while knowing what the ultimate solutions for those will be,
>> or
>>> we
>>>>> have solved everything, we'll be at the point where Oracle's SGA
>>>> (software
>>>>> grant agreement) can be worked on.
>>>>> 
>>>>> In other words, based on the above, the SGA would be executed as one
>> of
>>>> the
>>>>> LAST steps of the incubation period. After all, if we do uncover
>>>>> insurmountable issues during the incubation period, in particular in
>>>>> relation to licensing, having executed such a grant too early would
>>> lead
>>>> to
>>>>> a very difficult unwinding of the process.
>>>>> 
>>>>> In parallel to the licensing process described above, since we're
>>>> confident
>>>>> that in one way or another things will work out favorably, we could
>>>> decide
>>>>> to move the tutorials and other content from netbeans.org to the
>>> website
>>>>> structure, whatever that will be, at Apache, including setting up a
>>> Wiki
>>>>> structure in our new Confluence environment.
>>>>> 
>>>>> Comments to the above -- bring 'em on!
>>>>> 
>>>>> Gj
>>>>> 
>>>> 
>>> 
>> 

Craig L Russell
Secretary, Apache Software Foundation
[email protected] http://db.apache.org/jdo

Reply via email to