On 01/15/2013 10:03 PM, Luis Villa wrote:
On Tue, Jan 15, 2013 at 10:01 AM, Richard Fontana
<fontana(a)sharpeleven.org>
wrote:
> On 01/15/2013 11:02 AM, Luis Villa wrote:
>> Will the talk explain the purpose of c-l? :)
>
> Yes (though one response to the "What is the purpose" question is
"why
> does there have to be a purpose?").
I would say that it must have a purpose, because the cost of introducing a
new license is high; in particular, by creating more choices for everyone,
it means all participants in our ecosystem must learn something new and
face additional complexity every time they make a licensing decision.
This isn't to say that there should never be new licenses, just that
authors owe it to the broader community to consider these costs, and ask
the community to incur the costs only when there is a describable, positive
benefit to be gained. (AGPL and GPL v3, for example, would clear this bar-
they obviously addressed new, important issues. MPL 2 is a closer case,
though in its favor, the clearly-defined upgrade mechanism and explicit
deprecation by the license author mitigate some of the cost by replacing an
existing, older license.)
Sadly, I think this burden is higher for Serious Authors like you, Richard,
because your license is less likely to be ignored and therefore more likely
to impose costs on everyone.
That is quite flattering, but I do not think I deserve such flattery.
:-) Nevertheless, I understand what you are saying.
> If this is true, and I don't know whether it is, I assert
> that it implies that the degree of strong copyleft licensing,
> particularly for newer free software projects, is lower than it
> *ought* to be, as there are no viable alternatives to the GNU GPL
> family for developers otherwise potentially favorable towards strong
> copyleft policy.
> As a related point, I believe there is some continued commercial
> resistance to the GNU GPL license family (more so the v3 licenses than
> GPLv2, no doubt) and I believe this has some effect on upstream
> license use by projects that are not commercial as such, which goes
> beyond what one would expect if the only characteristic distinguishing
> copyleft from non-copyleft licenses were the minimum differential
> legal language necessary to implement the different legal policies of
> such licenses. I have encountered anecdotal evidence suggesting this
> is so.
Could you elaborate a bit on this? My sense is that the primary objection
for commercial users of the license has to do with the nature of the
copyleft rather than the specific language that implements the copyleft, so
for *commercial* adoption I think the drafting question is definitely only
marginal in effect. [Not to say a rewrite wouldn't help commercial
adoption, but that it is far from the primary issue.]
I agree it is not the primary issue. I would also say that the
commercial objection to copyleft specifically has to do with
(sometimes severely-mistaken) perceptions of FSF-influenced
interpretation of the GPL (which copyleft-next implicitly accepts as
essentially correct from a policy standpoint).
Nevertheless, it's at least an interesting secondary issue. As noted
it is more of an issue with the v3 family. Current copyleft-next
discards whole provisions of GPLv3: listing the more interesting ones
here:
* the anti-anti-circumvention provision
* the two Microsoft-Novell-deal-inspired provisions in GPLv3 section 11
* the "knowingly-relying" (aka "son of downstream shielding")
provision of GPLv3 section 11
* the reservation-of-implied-patent-licenses clause of GPLv3 section
11 (a recent deletion)
* the entirety of the anti-lockdown half of GPLv3 section 6
Also, more interesting from a GPLv2-oriented perspective,
copyleft-next goes further than GPLv3 in making the 'liberty-or-death'
clause merely a clarification of the 'no further restrictions' rule.
(Though as Bradley has interestingly noted, there is an argument that
copyleft-next has broadened liberty-or-death by not limiting it to a
condition on distribution. That goes into the still-nonexistent issue
tracker.)
All of the above provisions in my anecdotal-ish experience have met
with some commercial resistance (though really far less than one might
have expected). That itself isn't a reason to get rid of them in
copyleft-next (viewed as a sort of GPL fork), but what is a reason to
get rid of them is that such provisions don't actually accomplish
enough to justify whatever commercial friction exists -- the
anti-lockdown provisions *may* be a very important exception here
(though I'm not yet convinced of that). By constrast, while basic
FSF-interpretation strong copyleft also meets with commercial
resistance, this seems to be basically justified, the principal
problem being that FSF interpretation is not sufficiently well
understood by commercial vendors.
For the best example of what I'm getting at, consider a provision of
GPLv3 that was not deleted, but which was structurally transformed:
the patent license grant of GPLv3 section 11 paragraphs 1-3. I do not
think the underlying 'deep structure' of the patent license grants of
GPLv3 and copyleft-next are different, but I think their textual
implementations *look* dramatically different to the average
risk-averse corporate IP lawyer. Maybe the difference is still quite
marginal, though, so this point may well be smaller than I am making
it out to be.
> That's one long-winded partial answer to the "goal"
question. There
> are other goals, and other goals may yet emerge. At least one goal is
> personal (my personal feeling, which I can't shake, that GPLv3 is
> "unfinished", which I do not expect anyone else in the world to
> share).
I certainly sympathize :)
I thought you might understand what I mean. :)
- RF