Subject: Re: For Approval: Educational Community License 1.0
From: Christopher D. Coppola <chris.coppola@rsmart.com>
Date: Mon, 7 May 2007 08:18:19 -0700
Mon, 7 May 2007 08:18:19 -0700
Brian,

Thanks for engaging in this conversation and helping to think through  
the best solution. I hope you understand we've put a great deal of  
effort into making the progress we have. Though we agree the ECL 2.0  
proposal isn't a home run, we do think it's a significant improvement  
over the currently approved ECL 1.0 which we would end up having to  
stay on if 2.0 is not approved.

I'm curious about a comment you made...

> I just can't wrap my head around the idea that making the scope of the
> patent rights less clear for recipients (placing the burden on them to
> determine who the employers of contributors are, and whether the
> contributions were on the clock or off, etc) is an improvement that  
> merits
> certification as a new open source license -

It seems to me that with ECL 2.0 we're making the situation much more  
clear. Today it's anyone's guess whether or not any given patent held  
by the individual contributor or institution might be licensed. ECL  
1.0 makes no such claims and we're unsuccessful getting full  
compliance on the inbound (contribution agreement side because of  
this hang-up with many of the universities involved)... with ECL 2.0  
and the accompanying contribution agreements we make it clear on the  
outbound side (ECL 2.0) that patents are licensed for the scope of  
work intended to be contributed. We consider this a big step in the  
right direction. That said, we're eager to simply adopt the Apache  
license without modifications and we'll be working toward this as a  
next step.

Keep in mind that what we're requesting isn't a new license--it's a  
replacement, and it's really not intended for general use. Our  
license falls into the Special Purpose category because at present we  
clearly have some unique challenges as a community. That said, we're  
a very tight community. We do a good job of moving forward together.  
We have very clearly defined practices for license and IP management  
and a track record of very good performance using those practices to  
achieve great compliance on contributed and 3rd party code. All  
things considered, OSI approval of ECL 2.0 is a big step forward for  
our community, and for license proliferation. Without approval we'd  
necessarily need to stay with ECL 1.0 which negates the benefits of  
our diligent contribution agreement practices and does nothing for  
patent grants. There's also no good path from ECL 1.0 to a more  
popular license. We believe that ECL 2.0 (which is easily  
characterized as Apache 2.0 plus patent grant changes) gives us a  
path to adopting a popular license (Apache 2.0) in the future and  
eliminating ECL altogether.

Please consider our unique situation and the cohesiveness of our  
community in this request.

Thanks,
/Chris.



On May 1, 2007, at 10:29 AM, Brian Behlendorf wrote:

>
> Huh - one would think laws about IP created with state funds would  
> be even
> easier to work into this model, as it would presumably either assign
> ownership to the state, or specify that it be "public domain", or
> otherwise liberally licensed - in any case further separating the  
> academic
> institution's other IP (that they are interested in excluding from  
> this
> pool) from the open source project.
>
> I just can't wrap my head around the idea that making the scope of the
> patent rights less clear for recipients (placing the burden on them to
> determine who the employers of contributors are, and whether the
> contributions were on the clock or off, etc) is an improvement that  
> merits
> certification as a new open source license - even aside from the  
> lack of
> empathy I feel for IP holders who own so much IP they can't even be
> burdened to enumerate any that might apply to their contributions.  :)
>
> A license like this would seem to significantly weaken the value of  
> the
> software for new users - a new university wishing to use it would  
> now have
> greater worries about legal jeopardy from the organizations  
> sponsoring the
> work.  Given the precedent set and worries about the Blackboard  
> patent, I
> would have expected academic institutions lean towards greater  
> protections
> rather than fewer.
>
>         Brian
>
> On Mon, 30 Apr 2007, Christopher D. Coppola wrote:
> > Thanks for your thoughtful comments. We pursued this idea during  
> the Summit
> > we held on licensing among our community. Unfortunately we could  
> not get
> > agreement from some of the key members of our community (key  
> contributors).
> > Specifically it was cited that this approach would be  
> inconsistent with CA
> > law relating to ownership of IP created with state funds.
> >
> > /Chris.
> >
> >
> >
> > On Apr 30, 2007, at 12:58 PM, Brian Behlendorf wrote:
> >
> >> On Mon, 30 Apr 2007, Christopher D. Coppola wrote:
> >>> We had a chance to discuss the suggestion that we should  
> mention in the
> >>> ECL
> >>> 2.0 that there is only one sentence that is different from the  
> Apache 2.0
> >>> and
> >>> thatís all folks need to consider if they are already familiar  
> with the
> >>> Apache 2.0 license.  We think itís a very helpful suggestion,  
> and we are
> >>> submitting a revised draft of the license that makes the change.
> >>
> >> Now it seems like a tractable effort for me to add my 2c.  :)
> >>
> >> I think we can avoid needing to approve a new license here if the
> >> educational community in question is willing to make a change in  
> how it
> >> measures contributions to the project.  Here are the two lines  
> added to
> >> clause 3:
> >>
> >>> Any patent license granted hereby with respect to contributions  
> by an
> >>> individual employed by an institution or organization is  
> limited to
> >>> patent claims where the individual that is the author of the  
> Work is
> >>> also the inventor of the patent claims licensed, and where the
> >>> organization or institution has the right to grant such license  
> under
> >>> applicable grant and research funding agreements. No other  
> express or
> >>> implied licenses are granted.
> >>
> >> I am not a lawyer, this is not legal advice, but it does seem to  
> me like
> >> you're unnecessarily pushing a detail of whose patents and  
> copyrights
> >> you've collected on the front end (a purpose served by the various
> >> Contributor Agreements out there, like Apache's) into the license
> >> agreement on the back end.  It also seems like there may be
> >> employers/institutions who would be happy to fulfill the  
> obligations of a
> >> Contributor.
> >>
> >> Do this instead.  Within your project, for those employers who  
> can not go
> >> along with the obligations of being a Contributor, establish  
> that the
> >> Contributor is actually the individual, warranting on their own  
> that their
> >> contributions are their work and they can personally fulfill the
> >> obligations of contributorship.  In other words, take their  
> employers out
> >> of the equation.  That way their employers are not bound by the  
> patent
> >> commitment or any other.
> >>
> >> Since employment law usually acrues the IP ownership of works  
> created on
> >> employer (or on employer hardware) to the employer, what would  
> be required
> >> is a waiver or grant between employer and individual Contributor  
> that says
> >> that the Contributor owns the copyright and patent rights on  
> works they
> >> create on employer time.  The employer can then ask the employee  
> to grant
> >> back to the employer the unlimited rights of redistribution,  
> etc., that
> >> would normally exist if the employer was copyright holder.
> >>
> >> A Contributor doesn't even technically need to be the IP owner;  
> they just
> >> > need to have the required rights to grant.
> >>
> >> The qualitative difference to an end user is that even if the  
> project says
> >> "built with the help of developers working for University X",  
> there is no
> >> longer the assurance that they won't be sued down the road as  
> end-users by
> >> University X for something related to that very same code.   
> You'd owe it
> >> to your users to get very specific about who those Contributors  
> were -
> >> perhaps publicly publishing the list of who signed contributor  
> agreements.
> >>
> >> It's a repeal of patent self-defense a bit, but I guess  
> universities are
> >> used to that. I think this approach still protects the project  
> against
> >> submarined patents, too.
> >>
> >> Again, IANAL, but if this wouldn't work exactly, something close  
> to it
> >> must.
> >>
> >>        Brian
> >>
> >
>



Brian,

Thanks for engaging in this conversation and helping to think through the best solution. I hope you understand we've put a great deal of effort into making the progress we have. Though we agree the ECL 2.0 proposal isn't a home run, we do think it's a significant improvement over the currently approved ECL 1.0 which we would end up having to stay on if 2.0 is not approved.†

I'm curious about a comment you made...
I just can't wrap my head around the idea that making the scope of the
patent rights less clear for recipients (placing the burden on them to
determine who the employers of contributors are, and whether the
contributions were on the clock or off, etc) is an improvement that merits
certification as a new open source license -†

It seems to me that with ECL 2.0 we're making the situation much more clear. Today it's anyone's guess whether or not any given patent held by the individual contributor or institution might be licensed. ECL 1.0 makes no such claims and we're unsuccessful getting full compliance on the inbound (contribution agreement side because of this hang-up with many of the universities involved)... with ECL 2.0 and the accompanying contribution agreements we make it clear on the outbound side (ECL 2.0) that patents are licensed for the scope of work intended to be contributed. We consider this a big step in the right direction. That said, we're eager to simply adopt the Apache license without modifications and we'll be working toward this as a next step.†

Keep in mind that what we're requesting isn't a new license--it's a replacement, and it's really not intended for general use. Our license falls into the Special Purpose category because at present we clearly have some unique challenges as a community. That said, we're a very tight community. We do a good job of moving forward together. We have very clearly defined practices for license and IP management and a track record of very good performance using those practices to achieve great compliance on contributed and 3rd party code. All things considered, OSI approval of ECL 2.0 is a big step forward for our community, and for license proliferation. Without approval we'd necessarily need to stay with ECL 1.0 which negates the benefits of our diligent contribution agreement practices and does nothing for patent grants. There's also no good path from ECL 1.0 to a more popular license. We believe that ECL 2.0 (which is easily characterized as Apache 2.0 plus patent grant changes) gives us a path to adopting a popular license (Apache 2.0) in the future and eliminating ECL altogether.†

Please consider our unique situation and the cohesiveness of our community in this request.†

Thanks,
/Chris.



On May 1, 2007, at 10:29 AM, Brian Behlendorf wrote:


Huh - one would think laws about IP created with state funds would be even
easier to work into this model, as it would presumably either assign
ownership to the state, or specify that it be "public domain", or
otherwise liberally licensed - in any case further separating the academic
institution's other IP (that they are interested in excluding from this
pool) from the open source project.

I just can't wrap my head around the idea that making the scope of the
patent rights less clear for recipients (placing the burden on them to
determine who the employers of contributors are, and whether the
contributions were on the clock or off, etc) is an improvement that merits
certification as a new open source license - even aside from the lack of
empathy I feel for IP holders who own so much IP they can't even be
burdened to enumerate any that might apply to their contributions.† :)

A license like this would seem to significantly weaken the value of the
software for new users - a new university wishing to use it would now have
greater worries about legal jeopardy from the organizations sponsoring the
work.† Given the precedent set and worries about the Blackboard patent, I
would have expected academic institutions lean towards greater protections
rather than fewer.

††††††† Brian

On Mon, 30 Apr 2007, Christopher D. Coppola wrote:
> Thanks for your thoughtful comments. We pursued this idea during the Summit
> we held on licensing among our community. Unfortunately we could not get
> agreement from some of the key members of our community (key contributors).
> Specifically it was cited that this approach would be inconsistent with CA
> law relating to ownership of IP created with state funds.
>
> /Chris.
>
>
>
> On Apr 30, 2007, at 12:58 PM, Brian Behlendorf wrote:
>
>> On Mon, 30 Apr 2007, Christopher D. Coppola wrote:
>>> We had a chance to discuss the suggestion that we should mention in the
>>> ECL
>>> 2.0 that there is only one sentence that is different from the Apache 2.0
>>> and
>>> thatís all folks need to consider if they are already familiar with the
>>> Apache 2.0 license.† We think itís a very helpful suggestion, and we are
>>> submitting a revised draft of the license that makes the change.
>>
>> Now it seems like a tractable effort for me to add my 2c.† :)
>>
>> I think we can avoid needing to approve a new license here if the
>> educational community in question is willing to make a change in how it
>> measures contributions to the project.† Here are the two lines added to
>> clause 3:
>>
>>> Any patent license granted hereby with respect to contributions by an
>>> individual employed by an institution or organization is limited to
>>> patent claims where the individual that is the author of the Work is
>>> also the inventor of the patent claims licensed, and where the
>>> organization or institution has the right to grant such license under
>>> applicable grant and research funding agreements. No other express or
>>> implied licenses are granted.
>>
>> I am not a lawyer, this is not legal advice, but it does seem to me like
>> you're unnecessarily pushing a detail of whose patents and copyrights
>> you've collected on the front end (a purpose served by the various
>> Contributor Agreements out there, like Apache's) into the license
>> agreement on the back end.† It also seems like there may be
>> employers/institutions who would be happy to fulfill the obligations of a
>> Contributor.
>>
>> Do this instead.† Within your project, for those employers who can not go
>> along with the obligations of being a Contributor, establish that the
>> Contributor is actually the individual, warranting on their own that their
>> contributions are their work and they can personally fulfill the
>> obligations of contributorship.† In other words, take their employers out
>> of the equation.† That way their employers are not bound by the patent
>> commitment or any other.
>>
>> Since employment law usually acrues the IP ownership of works created on
>> employer (or on employer hardware) to the employer, what would be required
>> is a waiver or grant between employer and individual Contributor that says
>> that the Contributor owns the copyright and patent rights on works they
>> create on employer time.† The employer can then ask the employee to grant
>> back to the employer the unlimited rights of redistribution, etc., that
>> would normally exist if the employer was copyright holder.
>>
>> A Contributor doesn't even technically need to be the IP owner; they just
>> > need to have the required rights to grant.
>>
>> The qualitative difference to an end user is that even if the project says
>> "built with the help of developers working for University X", there is no
>> longer the assurance that they won't be sued down the road as end-users by
>> University X for something related to that very same code.† You'd owe it
>> to your users to get very specific about who those Contributors were -
>> perhaps publicly publishing the list of who signed contributor agreements.
>>
>> It's a repeal of patent self-defense a bit, but I guess universities are
>> used to that. I think this approach still protects the project against
>> submarined patents, too.
>>
>> Again, IANAL, but if this wouldn't work exactly, something close to it
>> must.
>>
>>††††††† Brian
>>
>