[License-discuss] [License-review] Encouraging discussion around the technicalities of licensing

[Marc Jones]

> > On Wed, Feb 6, 2019 at 10:40 PM Richard Fontana <
> richard.fontana at opensource.org> wrote:
> >>
> >>
> >> I lean towards disagreeing with this; I think that the business model
> of the license submitter can be a material consideration when assessing
> whether the proposed license meets the OSD and would guarantee software
> freedom.
> >
> >
> > Just the submitter's business model, or the business models of any
> possible user of the license?
>
> Just the submitter's, I think, if only because it may be too hard to
> think about all the possible ways an approved license might be gamed
> or (arguably) misused by a future licensor unconnected to the
> submitter. Also I think we ought to acknowledge that in practice
> licenses submitted by businesses tend not be used by other licensors
> (apart from forks). That was one of the reasons why the vanity
> corporate licenses of the earlier era of open source were problematic.
>
>
The reusability of the license and code seems to me be another way of
looking at the business model question.

The license text should be reusable by other persons. If the license is
only capable of being used by one company, what value is there in OSI
approving the license to anyone except the submitting company? This would
exclude licenses that hard code a specific person/company as being the
licensor and the code base being licensed. And I think it should also
exclude licenses that hard code the licensor being in an unreasonably
privileged position. It should also exclude licenses that required
significant changes or modifications to the license text to be used by
others. (One of the great advantages of FOSS licenses is that they are
capable of being used as reusable functions.) I know this is a deviation
from past requirements but the requirement that a license be reusable would
significantly reduce the risk of a return to vanity license proliferation.

The reusability of the code licensed under the license terms I think is
also a important consideration. If only the company sponsoring the license
is capable of complying with the license, the code is not really open
source. Or if in practice no company is capable of complying with the
license (including the license sponsor) then I don't see how anyone could
argue that it is actually open source. For instance if a license is
conditioned on a warranty clause that disclaims all liability even for
intentional wrongdoing that results in personal injury or death. No one can
offer a license under that condition in the United States (and I suspect in
most jurisdictions) since a warranty clause that broad would be void. Since
in practice no license is being offered, it should not be considered open
source.

OSI should be in the business of looking at the reusability of code from a
practical point of view. If using a particular license today is essentially
setting a licensing trap it should not be considered open source. OSI
should ask if, as a practical matter, the risk is high that that the
average developer or user using software would violate the license? If the
risk is high then the license should be rejected. For example if someone
were to draft a license that disallowed running the software on *both* a
GPL licensed kernels *and* on a BSD licensed kernels, as a practical matter
they have contributed nothing to the software commons and done nothing to
advance software freedom. The license should be rejected. I don't think it
would matter to me that you could run the software on a proprietary
licensed kernel or that you could find or write a MIT licensed kernel
capable of running it. I acknowledge that this argument would put OSI in
the position of not approving licenses today that it might have approved 20
years ago, or might still yet approve 20 years in the future. I think that
is unavoidable if the approval process is meant to have any practical
meaning. If OSI does not consider how the license would be used in the real
world. the approval process is likely to become academic, or useless.

To the extent the business model of the license sponsor informs us of how
practical it is to reuse the license or the code, I think the licensors
business model is relevant. If the license sponsor can not explain how
someone could reuse the licensor sponsor's code under a non-proprietary
business model, what use is there in the license? On the other hand it
might also be really obvious the other business models could apply, in
which case it seems silly to ask the license sponsor. But the mere fact
that the license sponsor is pursuing a proprietary business model is not
reason alone to reject the license. On the other hand it would be very
telling however if the license sponsor can't give an example of a situation
where the licensor sponsor would reuse code licensed under the license the
license sponsor is asking for approval for.

-Marc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20190211/16c20c90/attachment.html>