For Approval: Microsoft Permissive License

[Chris Travers]

On 9/9/07, Rick Moen <rick at linuxmafia.com> wrote:
>
> Quoting Donovan Hawkins (hawkins at cephira.com):
>
> > And given that background, do you see why the GPLv2 is _still_ better
> > than the GPLv3?
>
> This question completely lacks relevance to the legal issue at hand,
> which is a collective-work project leader's duty to avoid injuring
> contributors' economic and legal interests, when upgrading project
> licensing.


First, I agree that the GPLv2 *is* generally better than the GPLv3.  It is
simpler and easier to understand, and poses fewer ambiguities.


> If you were the project leader who took the Linux kernel to GPL v3,
> > Torvalds could claim damages in exactly the same way that taking it
> > non-copyleft would damage him.
>
> This is an ideological construing of the term 'damage' that would carry
> zero weight in tort law.


So, if I upgrade your code's license from the GPLv2 to the BSD, do you have
a cause of action?  Why or why not?  Note that unless you dual-license your
code, I have not caused you economic harm.  Or is injury broader than mere
damage to one's pocketbook (as CJ Roberts suggested in his conformation
hearing testimony).  While I don't think mere ideological issues would count
as injury, any time you make a substantive change to what a license
requires, you run the risk of hurting someone.  These changes do exist
between the GPL2 and the OSL,* between the GPL and the AGPL, and between the
GPL2 and the GPL3.  These are not safe license upgrades in the same sense of
MIT->BSDL.  Note that ESR does *not* say the GPL2 can safely be upgraded to
the GPL3.

*LeddgerSMB looked at the OSL at one point.  However, we decided that a
license which would require every public demo to distribute the source code
of that demo would ultimately make it harder to centralize and certify
official downloads, provide new security releases, and thus protect the user
of the software.  If the user of the software is not well protected, then
the project reputation as a whole suffers, and this does cause us harm,
which could possibly be quantified financially.

However, aside from the legal issues, if a license were to be upgraded
without *overwhelming* reason to do so, I think it would cost a project
dearly.

I ask you again:  If Tivo has any patches which have been accepted into the
kernel, would they have a cause of action from an upgrade to the license
which seems to target them?

> Perhaps other open source developers would be willing to follow your
> > advice if you'd like to indemnify him against a claim like this.
>
> Oh, give me a break.  If you want a warranty, Mr Hawkins, go buy one
> (from someone willing to deal with you, as I'm guessing you'd be a
> pretty high-maintence customer).  I'm just trying to describe the
> operation of tort law, and its ramifications for collective-work
> codebases, accurately.  If you prefer, instead, to fall back on the folk
> wisdom of law-ignorant coders, by all means go ahead.



I don't think you are quite accurate here.  Nor was I aware that you were
qualified to give legal advice.  (IANALE)


> > If you believe as Torvalds does and think GPL v3 is a very different
> > license following a different vision, you have every right to expect
> > your project leader to ask first.
>
> In _law_, actually, you do not.



How is Torvalds to know for certain what effect a license change would have
if he doesn't ask everyone involved?

Secondly, is it really wise to go on the line that "I can probably win in
court if this comes up?"  Nobody is likely to win in court in an open source
licensing dispute (unless you define winning as being vindicated in the
final judgement).  The only difference will be between those who lose and
those who lose more.

I guess the question you are asking is "Will I win in court if?"  While I
would suggest you ask, "Will this keep me out of court?"

> The fact that Torvalds never asked contributors to sign over copyright to
> > him suggests that he feels the same way.
>
> Non sequitur.



Not really.  The basic issue is this:

The GPL v3 has a number of restrictions in it which are not allowed under
the GPLv2.  We can all agree on that.  These restrictions (including
antitivoization and patent right termination on breaking the license) could
potentially harm copyright holders if implemented.  Torvalds has *no* way of
knowing this without getting the approval of every copyright holder or at
least asking them whether they are harmed or not.  Some could claim to be
harmed when not, just to block the license change.  Torvalds has better
things to do with his time than argue with these people in court.


Best Wishes,
Chris Travers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20070909/0405ef56/attachment.html>