Re: Fwd: That whole "Linux stealing our code" thing

[James Bruce <bruce@andrew.cmu.edu>]

Daniel Hazelton wrote:
> On Monday 03 September 2007 14:26:29 Krzysztof Halasa wrote:
>> Daniel Hazelton <dhazelton@enter.net> writes:
>>> The fact
>>> remains that the person making a work available under *ANY* form of
>>> copyright
>>> license has the right to revoke said grant of license to anyone.
>> Not after the licence has been given and accepted (and there might be
>> restrictions), unless of course the licence contained such reservation.
> 
> I hate to belabor the point, but you seem to be making the mistake of "The 
> license applies to the copyright holder" that I've seen a lot of people make 
> (and kept quiet about).

I believe you are making the mistake that the license on code has 
anything to do with what the author chooses to do in the future. 
Releasing something as BSD does not force the author to do anything in 
the future with his code, and he/she could add and relicence as he/she 
feels fit.  HOWEVER, that particular code has already been released as 
BSD, and the author no longer has control over that release.

> The person holding the copyright has all the legal standing to revoke a 
> license grant at any time. Licenses such as the GPL are not signed contracts, 
> and that means there are limits to what effect they can have on the copyright 
> holder.

I believe you are confusing the fact that an author can decide to 
release code under another license, with the existence of code under 
that earlier license.  The license grant comes from THE CODE (which 
bears a license), not THE AUTHOR.  I can use GPL code I get in the mail 
because the license on the work says I can do so, not because I 
contacted the author and got a specific grant.  If such a grant were 
only verbal, your theory might hold, but that doesn't apply to any OSS 
software under discussion here.

If your legal theory were true, I could sell you a book and then later 
demand that you destroy it.  I could also release something as public 
domain, and then later rescind that (I still hold the copyright on what 
I produced), and charge money from anyone who used it.  I think its safe 
to say that this does not happen in practice.  Please provide some 
examples to the contrary or caselaw if you want to convince me otherwise.

Furthermore, BSD/GPL software could not really exist under your legal 
theory; A programmer who wrote 30 year old core BSD code could wake up 
tomorrow and decide to require all BSD derivatives to remove his code or 
pay him for it (and the next day he could change the price again).  Open 
source software would not exist if such a liability were true, and 
companies like Sun could not be built up off of derivatives of it. 
Linux 0.01 is still available under a pre-GPL license if you can find a 
copy, and neither Linus (nor anyone else) can change that.

> If the license was of the "signed contract" type, or contained text stating 
> that the copyright holder was giving up all rights of revocation (etc...) I 
> could agree with you. As it stands, no "Open Source" license that I have seen 
> used on a major project contains any part that does that. In fact, the GPL is 
> the only license I can name (offhand) that even touches on the rights of the 
> copyright holder - and then it is in the form of "If you do X, Y or Z all 
> rights granted under this license are automatically revoked".
 >
> That is an "automatic clause" - not a limitation stating that the copyright 
> holder can only revoke under those conditions. The person holding the 
> copyright has quite a few rights - more than people believe - and not even 
> the most generous of Open Source licenses (except those that contain text 
> like "granted in perpetuity" or similar) even come close to being exempt from 
> the holder of the copyright not being able to summarily revoke a given 
> persons license.

There are plenty of rights, but retroactive changes to the license terms 
of something you've already distributed is not one of them.

  - Jim Bruce