Re: Software license?

On Fri, April 26, 2013 7:45 am, Dave Pawson wrote:
> On 25 April 2013 23:09, Tony Graham <tgraham@mentea.net> wrote:
>> I was looking at the "W3C Community Contributor License Agreement (CLA)"
>> [1] to see what Arved had got himself in for in providing the original
>> code and under what terms we were making the code public, but all it
>> says
>> is:
>>
>>    Any source code created by the Project is not subject to this
>>    CLA, but rather subject to separate licensing terms for that
>>    source code.
>>
>> Not very informative.
>>
>> Does that mean we need separate contributor agreements and license
>> assignments for software contributions?
>
> Not a lawyer, I'd interpret that as being 'what Arved attached' to the
> code?

Simply trying to use the same terminology as the agreement to which you
agreed when you joined the CG.

> Choose one, put a copy to the list and all use that?
>   GPL3 proposed.

GPL3 makes lawyers break out in hives.  Well, not exactly, but its
reputation precedes it and lots of people don't like its reputation.

Reviewing GNU's "How to choose a license for your own work" [1], I think
it helps to know what we want to achieve from having a license.  I'm with
Arved in wanting that people can use it but also wanting that they
acknowledge where it came from.  I don't think we want a copyleft license.
 I think we fit under the third paragraph of the 'Software' section where
it says 'The second is projects that implement free standards that are
competing against proprietary standards', and even GNU recommends against
copyleft for that situation.

It seems the three best alternatives are:

 - Apache License 2.0 [2]

 - Mozilla Public License (MPL) version 2.0 [3]

 - W3C Software Notice and License [4]

GNU blesses all of those as being GPL-compatible [5] and the OSI blesses
them all as being open source licenses [6].  The OSI classes Apache-2.0
and MPL-2.0 as popular and widely used, and classes the W3C license as
'non-reusable' because it's specific to the W3C, but that's not a problem
in this case.  GNU favours Apache-2.0 over other non-copyleft licences [7]
"since it blocks patent treachery".  About the only reason to go for the
MPL-2.0 license is if we extrapolate from the recent discussion of the
Saxon-CE event handling model to assuming that one day we'll end up also
adapting the Saxon-CE code itself.

Public domain isn't much of an option for us.  The OSI makes the point
that 'public domain' doesn't mean the same thing everywhere [8], and it it
was public domain, there couldn't be any requirement for acknowledging
where the software started from.  The public domain-like Creative Commons
"Zero" (CC0) license is not blessed by the OSI [10] because of the
non-transference of patent rights. GNU likes it all the same [9] while at
the same time warning about patent non-transference on another license
[11].

We don't actually know whether the W3C would want any software put in its
name, or whether there'd be additional paperwork for the legal niceties of
assigning copyright to the W3C.  I'll ask the CG support people if the W3C
has a policy about this.  If the W3C is happy to have random (from their
perspective) bits of code assigned to them, then I suggest we'll go for
the W3C license, otherwise I suggest we'll go for the Apache-2.0 license
and find some entity to be the copyright holder.

Regards,


Tony.

[1] http://www.gnu.org/licenses/license-recommendations.html
[2] http://directory.fsf.org/wiki/License:Apache2.0
[3] http://directory.fsf.org/wiki/License:MPLv2.0
[4] http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231
[5] http://www.gnu.org/philosophy/license-list.html#GPLCompatibleLicenses
[6] http://opensource.org/licenses/category
[7] http://www.gnu.org/philosophy/license-list.html#apache2
[8] http://opensource.org/faq#public-domain
[9] http://www.gnu.org/philosophy/license-list.html#CC0
[10] http://opensource.org/faq#cc-zero
[11] http://www.gnu.org/philosophy/license-list.html#clearbsd

Received on Friday, 26 April 2013 20:32:05 UTC