You are viewing a plain text version of this content. The canonical link for it is here.
Posted to legal-discuss@apache.org by Cliff Schmidt <cl...@apache.org> on 2006/06/01 00:44:53 UTC
Re: CCLA: "by combination" language
On May 30, 2006, at 10:57 AM, Doug Cutting wrote:
> Some folks seem to think that *any* patent owned by a contributor
> may eventually be licensed as it is subsequently combined with
> contributions of others. By this broad interpretation, any
> contribution to Apache under a CCLA could effectively freely
> license the contributor's entire patent portfolio to all users of
> Apache software. I don't this is the intent.
I thought I'd already explained the current policy to you, but I'll
try again:
It is not the case that *any* patent is eventually licensed -- only
those patent claims necessarily infringed by the Work due in part to
its inclusion of the Contribution. However, the "Work" does *not*
refer only to a single snapshot in time the moment when the
contribution was made, which would be of no practical use to the ASF
or its recipients. If Contributor A submits Contribution A to the
Work, and the next day Contributor B submits a patch to the same Work
-- any of Contributor A's dependent claims describing the features of
Contribution A when combined with features of the rest of the new
version of the Work, are also included in the grant. In simpler
terms, one cannot grant a license for their contributed framework
without also granting a license for the right to use that framework
with future contributions to the Work.
> Many companies with signed CCLAs are active Apache contributors and
> have invested in large patent portfolios that I doubt they intend
> to license freely to all users of Apache software.
This does not appear to be the case: BEA and IBM are clearly examples
of active Apache contributors invested in large patent portfolios.
You can re-read Jim and Jeff's posts for yourself, but it appears to
me that their interpretation is consistent with what I have described
above. As far as I know, their responses are the only responses to
this thread representing patent-owning companies contributing to
Apache. If others are out there with a different interpretation, I'd
certainly like to know!
<snip/>
> If a contributing company cannot determine which of their patents
> are freely licensed, they might be strongly discouraged from
> contributing.
IF this describes your situation:
- You (individual or your company) want to make a Contribution to
an ASF project; and
- you have, or plan to have, a patent describing features of an
invention that would necessarily be infringed by the potential
evolution of a project after accepting such Contribution; and
- such patent is so valuable that you do not want to offer a
royalty free license to recipients of the open source product.
THEN:
- I would suggest that you do not submit such Contribution to the ASF.
<snip/>
> What is the process for adding an entry to this? Can someone
> simply draft a new entry? Who would vote on such a change?
The process is that people or companies that disagree with the
current policy, as I have described it, are welcome to make their
cases on this list, as you have done. Should such feedback lead me
to believe that it is in the best interests of the ASF to change this
policy, I will certainly work on a more appropriate policy. In
addition, as an ASF member, you might want to see if there is a
consensus within the membership (or the board that represents the
membership) that this policy is actually not in the interests of the
Foundation. Should you find this to be the case, I'll definitely
reconsider this policy.
> An FAQ would clarify the intent, but companies would still be
> signing an admittedly vague CCLA. So long-term it seems that the
> CCLA itself should be clarified. What is the process for changing
> the CCLA?
An FAQ would definitely be helpful; I see folks like Henri are
already responding to some of these questions, which is great! I
also agree that the CCLA should be more clear about this and a couple
other issues; a revision of the CCLA and Software Grant is on my todo
list.
Cliff
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org
Re: CCLA: "by combination" language
Posted by Cliff Schmidt <cl...@apache.org>.
On May 31, 2006, at 7:32 PM, Roy T. Fielding wrote:
> I guess so, though quite frankly it doesn't matter what Cliff's
> interpretation of the clause may be. The only way to test it is
> to see what happens after someone who has signed the CCLA tries
> to sue the ASF (or a recipient) over a patent that they claim a
> given ASF work infringes. When that happens, the court is going to
> care about things like the opinion of the signer of the CCLA,
> the intent of the drafter of the CCLA, and the intent of the
> contributors. My opinion on the matter is a public record already,
> with a very long history behind it and a very specific use case
> excluded. As such, whatever our lawyers think, it is very unlikely
> that the CCLA can be imposed any more strictly upon the signatories
> than it was intended to be imposed by me. The CCLA will have to be
> changed if folks want to broaden its interpretation.
As the ASF's Vice President of Legal Affairs, I was directly asked
for clarification of the ASF's intent regarding the language used in
the patent license grant language in this agreement. I have answered
that question. I'm not going to argue about how much weight a court
would give to my answer. That's for the company who signs the CCLA
to consider.
Cliff
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org
Re: CCLA: "by combination" language
Posted by "Roy T. Fielding" <fi...@gbiv.com>.
On May 31, 2006, at 4:36 PM, Doug Cutting wrote:
> Cliff Schmidt wrote:
>> In simpler terms, one cannot grant a license for their
>> contributed framework without also granting a license for the
>> right to use that framework with future contributions to the Work.
>
> I'm not questioning how the "combination" affects what software the
> free license covers. I think it's reasonable for the free license
> to cover all future versions of the project. My question is
> limited to the determination of *which patents* are covered by the
> free license.
>
> Roy seemed quite clear on this:
>
> The clause is intended to exclude third-party
> contributions after the fact causing the infringement;
> that is the *only* reason we include the words "to which
> such Contribution(s) were submitted". Those words serve no
> other purpose than to limit the scope of the patents that
> infringe to claims that are potentially knowable by the
> contributor. Note that this limits the number of patents
> that infringe, not the number of Works to which the
> license is granted
>
> Do you differ with Roy on this point?
I guess so, though quite frankly it doesn't matter what Cliff's
interpretation of the clause may be. The only way to test it is
to see what happens after someone who has signed the CCLA tries
to sue the ASF (or a recipient) over a patent that they claim a
given ASF work infringes. When that happens, the court is going to
care about things like the opinion of the signer of the CCLA,
the intent of the drafter of the CCLA, and the intent of the
contributors. My opinion on the matter is a public record already,
with a very long history behind it and a very specific use case
excluded. As such, whatever our lawyers think, it is very unlikely
that the CCLA can be imposed any more strictly upon the signatories
than it was intended to be imposed by me. The CCLA will have to be
changed if folks want to broaden its interpretation.
Cliff, how we interpret the terms of a contract is not "policy".
Policy would be choosing which CCLA to require of contributors,
such as by changing the CCLA or issuing an addendum that is agreed
to as well. If the wording is unclear on its own, a court should
interpret the wording to the benefit of the contributor (against us).
> Why then are there equally large companies with large patent
> portfolios and active Apache contributors but no CCLA on file?
Because they have permitted their employees to contribute the company's
IP under the terms of the ICLA. They are well aware of that fact and
just as bound by it as they would be by signing a CCLA -- a corporation
cannot disclaim responsibility for the actions of its own employees
even when they aren't entirely aware of those actions, since the
corporation is inherently responsible for its own (lack of) oversight.
For example, if an employee of a company illegally dumps hazardous
waste generated by the company, both the employee and the company
are held responsible even if the company has a written policy
forbidding such illegal dumping. For similar reasons, a company
would not be able to successfully sue the ASF for making use of a
license given to the ASF by one of that company's own employees,
particularly when they are aware of such contributions and of
the purpose of the ASF.
....Roy
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org
Re: CCLA: "by combination" language
Posted by Jeffrey Thompson <jt...@us.ibm.com>.
Doug Cutting <cu...@apache.org> wrote on 06/02/2006 12:23:09 AM:
> Jeffrey Thompson wrote:
> > Doug Cutting <cu...@apache.org> wrote on 06/01/2006 11:53:21 AM:
> > > I don't follow this. The License is an agreement between the ASF
and
> > > users of Apache software, while the CCLA is an agreement between
the ASF
> > > and contributing companies. I don't see how the license alone can
> > > enforce the patent grant on a contributor.
> >
> > The contributor is submitting the material to Apache under the
license.
> > So, the license is also an agreement between the contributor and ASF
> > and the contributor and all users of Apache software.
>
> You're saying that by slapping the ASF license onto the contribution,
> the contributor grants a license to any patents implemented by it. And
> if the contributor is paid to do this, then his employer is implicitly
> granting patents that are infringed by the contribution. Is that right?
As Jim points out, there are situations where its clear that the
contributor is working on behalf of the employer, so the employer is
clearly making the grants. There's also situations where its clear that
the contributor is working on his/her own time and therefore the employer
is clearly not involved. And there's middle ground.
>
> So, by this logic, I see how the CCLA is not required. It provides a
> clearer paper trail, but doesn't actually change the situation much.
> The license alone is sufficient.
Right, one of the purposes of the CCLA is to make sure everyone agrees as
to what's going on. Eliminating any possible misunderstandings is a good
thing.
>
> This raises the question, if the license is sufficient, why is the ICLA
> required? Is this because a committer may contribute only a single line
> of code, which might not have the license attached to it, or might
> neglect to attach the license to a file, etc. Is that right, or is
> there some more compelling reason to require the ICLA?
>
Again, primarily (at least in my mind, with which others may disagree) it
is to eliminate possible misunderstandings. Its certainly possible to
contribute to an open source project without ever really looking at the
license. With the ICLA, you're asked to read and sign a document. Not a
hard thing to do, but it makes you pay attention to the license terms. The
other point is, while the grants are basically identical, there are a few
additional housekeeping clauses which wouldn't be relevant in the ASL2.0,
but make perfect sense in the ICLA (e.g., description of a process for
submitting third party code).
Jeff
Staff Counsel, IBM Corporation (914)766-1757 (tie)8-826 (fax) -8160
(notes) jthom@ibmus (internet) jthom@us.ibm.com (home) jeff@beff.net
(web) http://www.beff.net/
Re: CCLA: "by combination" language
Posted by Doug Cutting <cu...@apache.org>.
Jeffrey Thompson wrote:
> Doug Cutting <cu...@apache.org> wrote on 06/01/2006 11:53:21 AM:
> > I don't follow this. The License is an agreement between the ASF and
> > users of Apache software, while the CCLA is an agreement between the ASF
> > and contributing companies. I don't see how the license alone can
> > enforce the patent grant on a contributor.
>
> The contributor is submitting the material to Apache under the license.
> So, the license is also an agreement between the contributor and ASF
> and the contributor and all users of Apache software.
You're saying that by slapping the ASF license onto the contribution,
the contributor grants a license to any patents implemented by it. And
if the contributor is paid to do this, then his employer is implicitly
granting patents that are infringed by the contribution. Is that right?
So, by this logic, I see how the CCLA is not required. It provides a
clearer paper trail, but doesn't actually change the situation much.
The license alone is sufficient.
This raises the question, if the license is sufficient, why is the ICLA
required? Is this because a committer may contribute only a single line
of code, which might not have the license attached to it, or might
neglect to attach the license to a file, etc. Is that right, or is
there some more compelling reason to require the ICLA?
Thanks,
Doug
[These are my opinions, not those of my employer.]
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org
Re: CCLA: "by combination" language
Posted by Jeffrey Thompson <jt...@us.ibm.com>.
Doug Cutting <cu...@apache.org> wrote on 06/01/2006 11:53:21 AM:
> Jeffrey Thompson wrote:
> > I hesitate to guess at the answer to that question, however, I'd like
to
> > point out that it is probably NOT the language of the patent grant
> > that's the cause. The patent language in the CCLA is identical to the
> > language in the Apache License 2.0, which applies to the Corporation
> > through the authorized Contributions by its employees. Refusing to
sign
> > the CCLA doesn't negate that patent grant.
>
> I don't follow this. The License is an agreement between the ASF and
> users of Apache software, while the CCLA is an agreement between the ASF
> and contributing companies. I don't see how the license alone can
> enforce the patent grant on a contributor.
The contributor is submitting the material to Apache under the license.
So, the license is also an agreement between the contributor and ASF and
the contributor and all users of Apache software.
> The iCLA is an agreement
> between an individual contributor and the ASF, and hence, to my
> understanding, only implicates patents owned by that individual.
Maybe.
> So I
> still don't see how the ASF can grant licenses to patents owned by
> companies that have not signed a CCLA. Can someone please explain this
> to me? What am I missing?
If the individual making the contributions is an employee of a company and
is participating in Apache at the request of the employer and/or
representing the employer's interests and/or contributing materials owned
by the employer with the consent of the employer, its clear that the
employer is the real contributor and is bound by the Apache license. The
definitions in the license itself makes that point clear, though normal
agency law would come to the same conclusion w/o that language.
On the other hand, if the employee is participating on his/her own time
and the participation has nothing to do with his/her employment and the
employer, by force of law or employment agreement, has no claim on the
code written by the employee, then the employer is not involved in any
way, doesn't have to make any grants at all, and you're position would be
right.
>
> Also, the patent language in the CCLA is *not* identical to the language
> in the license. At least one difference is that the license uses the
> phrase, "by combination of their Contribution(s) with the Work to which
> such Contribution(s) *was* submitted" (emphasis added). The license
> thus seems to imply more strongly the intended reading, where the
> evaluation of which patents are licenseable is to be made against the
> Work at the time of contribution, not at a later date.
So, you're saying that "were submitted" more strongly implies an intent to
place a temporal restriction on the scope of the Work than "was
submitted"?
Is it possible that someone in one case thought that "Contribution(s)",
being of undefined plurality, should be thought of as singular and used
"was" and that someone else when looking at the CCLA thought that
"Contribution(s)" should be thought of as plural and corrected the mistake
to "were".
Jeff
>
> Thanks,
>
> Doug
Staff Counsel, IBM Corporation (914)766-1757 (tie)8-826 (fax) -8160
(notes) jthom@ibmus (internet) jthom@us.ibm.com (home) jeff@beff.net
(web) http://www.beff.net/
Re: CCLA: "by combination" language
Posted by Doug Cutting <cu...@apache.org>.
Jeffrey Thompson wrote:
> > Why then are there equally large companies with large patent portfolios
> > and active Apache contributors but no CCLA on file?
>
> I hesitate to guess at the answer to that question, however, I'd like to
> point out that it is probably NOT the language of the patent grant
> that's the cause. The patent language in the CCLA is identical to the
> language in the Apache License 2.0, which applies to the Corporation
> through the authorized Contributions by its employees. Refusing to sign
> the CCLA doesn't negate that patent grant.
I don't follow this. The License is an agreement between the ASF and
users of Apache software, while the CCLA is an agreement between the ASF
and contributing companies. I don't see how the license alone can
enforce the patent grant on a contributor. The iCLA is an agreement
between an individual contributor and the ASF, and hence, to my
understanding, only implicates patents owned by that individual. So I
still don't see how the ASF can grant licenses to patents owned by
companies that have not signed a CCLA. Can someone please explain this
to me? What am I missing?
Also, the patent language in the CCLA is *not* identical to the language
in the license. At least one difference is that the license uses the
phrase, "by combination of their Contribution(s) with the Work to which
such Contribution(s) *was* submitted" (emphasis added). The license
thus seems to imply more strongly the intended reading, where the
evaluation of which patents are licenseable is to be made against the
Work at the time of contribution, not at a later date.
Thanks,
Doug
[These statements are my own, not those of my employer.]
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org
Re: CCLA: "by combination" language
Posted by Jeffrey Thompson <jt...@us.ibm.com>.
Doug Cutting <cu...@apache.org> wrote on 05/31/2006 07:36:26 PM:
> Cliff Schmidt wrote:
> >
> > This does not appear to be the case: BEA and IBM are clearly examples
> > of active Apache contributors invested in large patent portfolios.
>
> Why then are there equally large companies with large patent portfolios
> and active Apache contributors but no CCLA on file?
>
I hesitate to guess at the answer to that question, however, I'd like to
point out that it is probably NOT the language of the patent grant that's
the cause. The patent language in the CCLA is identical to the language
in the Apache License 2.0, which applies to the Corporation through the
authorized Contributions by its employees. Refusing to sign the CCLA
doesn't negate that patent grant.
Jeff
Staff Counsel, IBM Corporation (914)766-1757 (tie)8-826 (fax) -8160
(notes) jthom@ibmus (internet) jthom@us.ibm.com (home) jeff@beff.net
(web) http://www.beff.net/
Re: CCLA: "by combination" language
Posted by Doug Cutting <cu...@apache.org>.
Cliff Schmidt wrote:
> In simpler terms,
> one cannot grant a license for their contributed framework without also
> granting a license for the right to use that framework with future
> contributions to the Work.
I'm not questioning how the "combination" affects what software the free
license covers. I think it's reasonable for the free license to cover
all future versions of the project. My question is limited to the
determination of *which patents* are covered by the free license.
Roy seemed quite clear on this:
The clause is intended to exclude third-party
contributions after the fact causing the infringement;
that is the *only* reason we include the words "to which
such Contribution(s) were submitted". Those words serve no
other purpose than to limit the scope of the patents that
infringe to claims that are potentially knowable by the
contributor. Note that this limits the number of patents
that infringe, not the number of Works to which the
license is granted
Do you differ with Roy on this point?
>> Many companies with signed CCLAs are active Apache contributors and
>> have invested in large patent portfolios that I doubt they intend to
>> license freely to all users of Apache software.
>
> This does not appear to be the case: BEA and IBM are clearly examples
> of active Apache contributors invested in large patent portfolios.
Why then are there equally large companies with large patent portfolios
and active Apache contributors but no CCLA on file?
> IF this describes your situation:
> - You (individual or your company) want to make a Contribution to an
> ASF project; and
> - you have, or plan to have, a patent describing features of an
> invention that would necessarily be infringed by the potential
> evolution of a project after accepting such Contribution; and
> - such patent is so valuable that you do not want to offer a royalty
> free license to recipients of the open source product.
>
> THEN:
> - I would suggest that you do not submit such Contribution to the ASF.
The problem with this logic is that the "potential evolution of a
project after accepting such Contribution" is not knowable. So one
cannot in fact evaluate your formula.
>> What is the process for adding an entry to this? Can someone simply
>> draft a new entry? Who would vote on such a change?
>
> The process is that people or companies that disagree with the current
> policy, as I have described it, are welcome to make their cases on this
> list, as you have done. Should such feedback lead me to believe that
> it is in the best interests of the ASF to change this policy, I will
> certainly work on a more appropriate policy. In addition, as an ASF
> member, you might want to see if there is a consensus within the
> membership (or the board that represents the membership) that this
> policy is actually not in the interests of the Foundation. Should you
> find this to be the case, I'll definitely reconsider this policy.
So if I am unable to convince you then I could call for a vote on the
board@ list, and any decision there must be a consensus, not majority,
decision? (Most actions of the board and members seem to be majority,
not consensus based, aren't they?)
Doug
---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only. Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF. See <http://www.apache.org/licenses/> for
official ASF policies and documents.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org