[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: CVS update: subversion/subversion/libsvn_client add.c apply_edits.c checkout.c client.h commit.c delete.c status.c update.c

From: Greg Stein <gstein_at_lyra.org>
Date: 2000-11-22 22:24:39 CET

Not true. Most of those phrases say that you can release the code under the
later license. In essence, you can make all of your changes and then release
under Vn+1, thus disallowing people to use Vn for your changes.

You're right that it doesn't affect *existing* developers/users. It affects
changed/later versions of the software.

Go back to the AOL/Netscape scenario. Let's say that they change the license
to permit no copying and to no longer be Open Source. Now, let's say that
they change a couple things in the browser: 1) eliminate some kind of
compatibility [to create locking], and 2) add some really cool feature that
users want. Now you have a scenario where their marketing engine is going to
flood the market with a lockin-browser, and you have to duplicate their work
to get the cool feature. And you have no source to do it. And you're a major
contributor to Mozilla, and they went and did this with *your* work.

Wouldn't that piss you off?

Personally, I like to contribute to software under a license that I know
will apply to everybody. I don't want that license to magically change
underneath me, to something that I disagree with.

Now, I recognize most of this is philosophical because it is all presumptive
on CollabNet issuing a later version of the license that is draconian. But
the philosophy is still important [to me]. Consider the evil case where
CollabNet gets sucked up by their big buddy Sun. Sun says "wow, everybody on
the 'Net is using SVN. Let's take advantage of that." They issue a new
license with the evil form. Then they start updating and releasing new
copies of SVN. Fine, no problem, you say... we just fork. But when we fork,
we're still doing it with the "or later version" phrase in there! Sun could
continue to take our work, integrate it into theirs, and release it under
their V2 Draconian license.

The counter case is if the license said "V1 only". We could fork and develop
under the old license. Sun could not take our work and put it into their V2
licensed software.

Again, this is philosophy. But those darn licenses are philosophy to start
with. They haven't and probably won't ever get tested in court. But if we're
going to put pure philosophy into our code, then let's do it right.

Cheers,
-g

On Wed, Nov 22, 2000 at 12:55:01PM -0600, Karl Fogel wrote:
> Greg, those retroactive later-version licenses don't work the way
> you're saying.
>
> Once you've released code under a particular license, no one can take
> those terms away. However, you can give people the *option* of using
> your code under a different license, or not (that's the "this version
> X, or any later version" clause).
>
> Thus, people will always have whatever rights you originally released
> the code under -- CollabNet cannot change that later. However, you'd
> be explicitly granting them the right to *choose*, later on, to use
> some other terms. Since our current license already permits
> proprietary forks, this can't result in anything too awful. But it
> does give CollabNet the ability to retroactively rerelease things
> under even more liberal terms in the future.
>
> The scenarios you describe below don't happen:
>
> Greg Stein <gstein@lyra.org> writes:
> > As a user of software, I really dislike that "any later version" phrase.
> >
> > *) consider the NPL/MPL: it has this "feature". Are you happy contributing
> > to that project, knowing that AOL/Netscape can release a new NPL/MPL that
> > says "nobody can copy this software without paying us our licensing fee"
> > and attach it to the code you wrote? If you were a user, would you
> > appreciate new licensing that said you must pay for using Mozilla?
>
> But that's not what happens, because you still always have the option
> of choosing the earlier (original) terms. In fact, you (the user) get
> to select whichever version of the license you like the most, from all
> the versions since the earliest, inclusive.
>
> > *) consider the GPL: are you happy to contribute to that, knowing that RMS
> > could release GPL V3 that states that you can only use the software on
> > systems that contain zero proprietary software? He could do it, and
> > software authors could elect to use V3 and there isn't much you can do
> > about it since it *said* they could do that.
>
> Same rebuttal.
>
> > *) Linus removed that "any later version" phrase from (his portions of)
> > Linux because he wasn't sure what RMS was going to do in V3, and he
> > didn't want those changes to magically apply to his software.
>
> Same.
>
> > I say that software should be released under a specific license. If you want
> > to change the license, then re-release the software with the change. Let the
> > people decide whether they want to use the new or old license.
> >
> > But this whole "hey, we can change it" is a bit scary for developers and
> > users of the software.
>
> Not scary if you realize all the options available to every user.
>
> -K

-- 
Greg Stein, http://www.lyra.org/
Received on Sat Oct 21 14:36:15 2006

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.