[License-review] Submission of the Upstream Compatibility License v1.0 (UCL-1.0) for approval

[Henrik Ingo]

Nigel

Thank you for your extensive explanation of your motivations.

It seems this license makes sense to you, and there could be some
government sponsored code that could be open sourced under it. I also
think it is trivially the case that this should be accepted by the
open source community, since for the downstream users it is
essentially Apache licensed and they could drop the UCL license if
they choose to do so.

Just for the sake of clarity: You do realize, that a recipient of UCL
software could create a derivative work, drop the UCL license, and
just use the Apache license. Another recipient of that derivative work
could then take the entirety of the code, now licensed under Apache
license, and add new features that would remain closed source. The
Apache license allows this. So this second recipient of the code is no
longer obligated to share his derivative work with the creator of the
original work (you).

henrik



On Tue, Feb 28, 2017 at 12:03 AM, Nigel T <nigel.2048 at gmail.com> wrote:
> Henrik,
>
> Nope :). Still the objective.
>
> The Original Work is not dual licensed but exists only as UCL.  All
> derivatives of the original work must be apache.  Should this be
> explicitly spelled out in the license text itself or just the
> supporting material?
>
> If this doesn't apply to derivative works of derivative works then
> that is an issue with this change.
>
> We have built spectral signature software for the US government that I
> would like to open source.  UCL provides that middle ground where I
> can state to the stakeholders that any community built enhancements
> would be able to be incorporated into the original work that would
> remain closed source GOTS and potentially a commercially supported
> product (which we don't do that very often).
>
> Keeping the code for a commercial product closed source has been the
> default option in the past.  There exists similar software, funded by
> other governments (and their citizens), that is the basis for
> expensive closed source products. I would like to change that.
>
> There are pieces of the original work that won't be released under UCL
> that is not freely distributed to entities outside the USG.  GPL isn't
> a good alternative for the whole project as incorporating community
> built GPL enhancements would require the distribution of the original
> source code to anyone we give the binaries to that isn't part of the
> federal government.
>
> There are reasons the original sponsoring agency would not want to
> distribute source.  The code is all unclassified but there are
> portions that may be FOUO (For Official Use Only).  That requires that
> the source is encrypted at rest and we can't distribute FOUO material
> to foreign entities (at least under the guidelines we follow).
> Compiled binaries can be decompiled but any FOUO comments in the
> source won't appear.  How we do X isn't likely sensitive. Why we do X
> may be.  So simple code obfuscation is good enough.
>
> Under UCL though we can give the bulk of the code away under a
> copyleft and any enhancements made in the commercial or academic world
> can be folded back into the original GOTS system without licensing
> concerns or need to manage a foundation or a CLA scheme.
>
> I just don't want to convince my sponsor that open sourcing is a good
> idea and then later on have to explain why we can't use stuff built on
> what the government already paid for because the derivative work was
> licensed under an incompatible (for his needs) license.
>
> If all derivative works (including their derivatives) of the original
> work must be apache then the original sponsor can use it.
>
> Regards,
>
> Nigel
>
>
> Sent from my iPhone
>> On Feb 24, 2017, at 3:18 AM, Henrik Ingo <henrik.ingo at avoinelama.fi> wrote:
>>
>> Hi Nigel
>>
>> A couple questions about this revised UCL license:
>>
>> 1. Your original rationale for this submission was "To provide
>> developers wishing to open source an existing closed source product
>> with an Open Source license that is Copyleft for downstream developers
>> and Permissive for upstream developers".
>>
>> Have you abandoned this goal? It seems to me this revised license is
>> essentially Apache 2.0 for both upstream and downstream developers?
>> (Recipients of derivative works could choose to drop UCL at any time,
>> otoh it is not possible to drop the Apache license option, since it is
>> hard coded into the UCL.)
>>
>> 2. Given this, why wouldn't a creator of an Original Work simply use
>> Apache from the beginnng?
>>
>> 3. Sorry if this was already answered earlier, but: Are you working on
>> behalf of someone that is looking to license some currently closed
>> source software under this license? If yes, can you share who that is,
>> and what the software in question is or could be?
>>
>> henrik



-- 
henrik.ingo at avoinelama.fi
+358-40-5697354        skype: henrik.ingo            irc: hingo
www.openlife.cc

My LinkedIn profile: http://fi.linkedin.com/pub/henrik-ingo/3/232/8a7