[OpenID] OpenID license

Wed Jul 28 13:21:54 UTC 2010

On 23 July 2010 23:27, David Recordon <recordond at gmail.com> wrote:

> Inline...
>
> On Fri, Jul 23, 2010 at 12:33 PM, Nathan <nathan at webr3.org> wrote:
>
>> To summarise then (with regards OpenID/OIDF specifications):
>>
>> All contributors have signed patent non assert agreements
>>
>
> Yes.
>
>
> The signed agreements can (not) be found here http://openid.net/ipr/
>
>
> This directory listing works again. It contains the non-asserts for OpenID
> 2.0 from those contributors. It doesn't look like the contribution
> agreements from PAPE are currently listed online, but every contirbutor
> signed one (and they should be put online).
>
>
> No patents from contributors covering OpenID specifications are disclosed.
>>
>> The non assert agreements protect contributors from each other (to an
>> extent), they do not protect implementers.
>>
>
> Incorrect. The non-asserts cover implementors as well.
>
>
> As far as you know the OpenID specs are not patent encumbered, but you
>> advise implementers to access the legal situation with their legal council
>> before writing a line of code, and if worried to go and license the relevant
>> patent(s) or get patent non asserts.
>>
>
> As is true for implementing any specification you find on the web.
>
>
> The patent(s) may cover parts of other existing or future protocol
>> specifications from non associated third parties, and of course the
>> implementations of those.
>>
>> The copyright on OpenID specifications mean they cannot be released under
>> CC-zero licenses (or similar), the licenses which are compatible are
>> unknown, Janrain has opted for Apache V2 but may still be infringing on
>> patents (as all implementations may be).
>>
>
> These issues are orthogonal. The contribution agreements include a
> copyright license which allows the Foundation to distribute the
> specifications. We did not choose to adopt a CC license as the included
> copyright license can be shorter.
>
> Implementations, such as JanRain, can choose their own licensing terms for
> their implementations. This is completely separate from the licensing of the
> specifications.
>
>
> The 'OIDF hereby disclaims any responsibility for identifying the
>> existence, or for evaluating the applicability, of any patents, patent
>> applications, or other rights (including copyrights) claimed to be
>> applicable to any Specification and will take no position on the validity or
>> scope of any such rights.'
>>
>> The general advice is that because of the legal costs of a patent
>> infringement case it's likely that anybody implementing will be infringing
>> patents (if there are any, but they aren't disclosed) but they most likely
>> won't be sued because of the costs involved.
>>
>> So, do I take it that I should just get on and implement the
>> specifications, go for a license which keeps all rights reserved to the OIDF
>> and hope for the best; ignore the patent matter, and if manages to get a
>> business to the value where patent infringements would be worth going after
>> seek legal council and worry about it then.
>>
>
> I'm not a lawyer, can't given you legal advice, and don't understand your
> business. That said, hundreds of people and companies have implemented
> OpenID around the world.
>

I think the story here is that, "the glass is 99% full".  Having followed
this for quite some time, I think it's nothing short of remarkable the
achievement made by the members in advocating an open web.  I remember in
the days of passport something like OpenID just seemed unimaginable.

However, so great has the momentum of this movement been, that I think the
conversation has shifted, in as little as two years, I would say "Open" has
from being the exception to being the norm.  This is a trend that seems to
be accelerating, if anything.

I would argue that from a business point of view, in the current environment
the intangible value to a business of the very remote chance of using
patents should be weighed against the very real risk that an (even slightly)
restrictive licensee, could be viewed negatively.

I think the classic case in point is the evolution of the web itself.  In
the early 90s gopher was the predominant protocol.   It was TimBL persuading
CERN to license it royalty free vs gopher's fear of being a restrictive
license that was the clincher in evolution of both systems.  (Even though
gopher got it right in the end, but also too late!) [1]

Personally I dont mind too much which choice is taken, as I have a wealth of
options to implement.  But OpenID is a tech I like, one I'd want to promote
and implement, and also see grow in success.

I'm not an OIDF member, but would now be a reasonable time to suggest, it's
may be everyone's interests to consider a royalty free license?

[1] In February 1993, the University of
Minnesota<http://en.wikipedia.org/wiki/University_of_Minnesota>announced
that it would charge licensing fees for the use of its
implementation of the Gopher
server.[3]<http://en.wikipedia.org/wiki/Gopher_%28protocol%29#cite_note-2>As
a consequence of this, some users were concerned that a licensing fee
would also be charged for independent
implementations.[4]<http://en.wikipedia.org/wiki/Gopher_%28protocol%29#cite_note-3>
[5] <http://en.wikipedia.org/wiki/Gopher_%28protocol%29#cite_note-4> In
contrast, no such limitation has ever been imposed on the World Wide Web.
The University of Minnesota later re-licensed its Gopher software under the GNU
GPL <http://en.wikipedia.org/wiki/GNU_GPL>.[6]<http://en.wikipedia.org/wiki/Gopher_%28protocol%29#cite_note-5>

http://en.wikipedia.org/wiki/Gopher_%28protocol%29

>
>
> Am I correct?
>>
>> Best,
>>
>> Nathan
>>
>>
>> Chris Messina wrote:
>>
>>> On Wed, Jul 21, 2010 at 6:48 PM, Nathan <nathan at webr3.org> wrote:
>>>
>>>  Essentially the non-asserts are about protecting the creators of the
>>>>> technology, and less about protecting the implementors. It's up to each
>>>>> implementor to assess the legal situation with their own counsel (if
>>>>> it's
>>>>> important to them) before writing a line of code. The contributors
>>>>> obviously
>>>>> can't do that for you, they can only assess their own legal situation
>>>>> and
>>>>> act according to their interests.
>>>>>
>>>>>  well I can't afford to do that, nor do I have the time so doesn't
>>>> really
>>>> leave me much choice I guess :(
>>>>
>>>
>>>
>>> Most people can't afford this (including me, personally) and implement
>>> anyway.
>>>
>>> It's up to you, as I said, to determine your risk and proceed
>>> accordingly.
>>>
>>> If you can't or won't implement OpenID because you're concerned about
>>> being
>>> sued for patent infringement, consider how much patent litigation costs
>>> and
>>> then weigh that against the likelihood that anyone would really go after
>>> anyone worth less than 10s of millions of dollars for patent
>>> infringement.
>>>
>>> Hell, if anyone is really worried about your implementation, you can
>>> always
>>> go license the relevant patent(s).
>>>
>>>
>>>  Not today. Depends on the copyright license that applies. The default is
>>>>> all
>>>>> rights reserved, so until we specify otherwise, that's the doctrine
>>>>> that
>>>>> applies.
>>>>>
>>>>>  okay, assuming that Apache License V2.0 is compatible given that
>>>> janrain
>>>> openid implementations are released under it, any word on CC
>>>> Attribution-ShareAlike (for an implementation).
>>>>
>>>>
>>> Copyright license on code is separate from patent licenses. Janrain
>>> libraries could still infringe patents, but you could at least create
>>> derivative works or fork the libraries thanks to the copyright license.
>>>
>>> Just remember to keep those issues separate.
>>>
>>> Chris
>>>
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20100728/4818ba26/attachment.html>