RE: Clear up my License information

> Patrik Stridvall wrote:
> 
> > > I think it's *still* unclear.  Easy to show this in java: Let's
> > > say you have a LGPLed jar file in package "gnu.foo.bar".  
> I believe
> > > from what rms has said in the past, if you write "package 
> gnu.foo.bar;" in
> > > your code, I think a LGPL claim can be made on what you
> > > write, especially if you call something in gnu.foo.bar.
> 
> > LGPL primary claims what the LGPL:ed code calls so I guess 
> you mean the GPL.
> 
> LGPL also makes association claims ("work as a whole") - I 
> said what I meant.

OK, but it unlike the GPL contains exceptions for users of the
LGPL:ed code so I can't see why this is relevant.
  
> > In any case. Regardless of what licence the patch as a 
> whole is (or must be
> > under) doesn't mean that it isn't possible to extract X11 
> license code from
> > that patch.
> 
> Hard to say.  If you take rms' view of how viral the GPL 
> derived clause is, then
> I'd say it is possible that you couldn't.  Remember that any 
> claim is not going
> to be made by the person who made the (X11) patch, but by 
> some other 3rd party -
> and could claim *all* the patch is derivative code.

First of all the (L)GPL have already been fullfill since X11 license
is compatible with (L)GPL, so get got what they asked for in the license.

So any derivation claim can't be based on anything the (L)GPL says.
They have to claim derivation in a general sense. This is I think
in the case of Rewind is all but impossible.

> > Not even the GPL claim that code must be ONLY under the GPL.
> > It must be under at least one GPL compatible license but it 
> says nothing
> > about other licenses.
> 
> This is incorrect.  Patches on GPLed work MUST be GPLed (not 
> GPL compatible).
> You may relicense it as something else also, but at least one 
> MUST be GPL.

If you have released the code under an GPL:ed compatible license
(read: on that have less restrictions that the GPL) you can't
truly release it under the GPL since you already have given
up some of the orginal rights that copyright law grants you.

In short:
You can consent (or not) to its release as GPL but it has no legal
meaning since, as I said, you already have giving up some of your rights.

If the GPL had insisted on this the LGPL => GPL paragraph would have
been meaningless for starters. It doesn't so that is that.

> > In short:
> > The GPL claim doesn't (and can't) increase the restrictions 
> of the code
> > published by other authors regardless of whether they use 
> GPL:ed code or
> > not. The restrictions or lack of the restrictions will alway remain.
> 
> Probably, but think of this case:  Let's use wine as an 
> example.  Let's say I
> took the LGPL wine when it was first converted to LGPL (so 
> wine == rewind) and
> I publically stated that I took LGPLwine.tar.gz and made a 
> propriatary change
> to it.  Could I get sued (assume that both trees are 
> identical for the sake of
> argument)?

Well, you could always get sued right or wrong. :-)

Seriously, anybody suing you would have no possibillity of winning.
The only thing the a license change can do to old code is changing
the burden of proof. Nothing else. It can't make anybody more or
less guilty (or liable).
  
> > > one could come up with an if (let's say if( setup_fn() &&
> > > setup_fn1() ) ).
> > > It can be a gray area.
> > 
> > No, it can't since the patch wouldn't apply, so manual 
> "surgery" is needed.
> 
> Patch doesn't do changes?  

Patch doesn't do any changes for parts (hunks) that doesn't apply.
This would be the case for the change you suggested above. 
So merge would need to be done manually.

> > Regardles on what the merger actually did it would be 
> almost impossible to
> > prove that an actual infringement have taken place unless 
> the substantial parts
> > was almost exactly the same save whitespace or variable 
> substitutions.
> 
> This has yet to be seen.  I think this is where the 
> "contamination" would
> fit in, and I believe there are legal precidents that argue 
> against you here
> (the IBM bios case comes to mind).

IIRC the possible contamination was because the IBM BIOS
contained tradesecrets. This have nothing to do with the
fact that IBM held copyright to the BIOS.

Mere copyright can't cause contamination. Period.
If you have published it you have published it and
all the ideas, facts, methods etc published are free
for taking.

Obvious tradesecret status is lost if you publish it.
However IIRC the issue relevent the IBM case is whether the
possible tradesecrets (ideas, fact, methods etc) really was
published when it was released as a ROM chip containing machine
code language not meant for and unsuitable for human reading.

This is area of law not well covered in the courts.
Especially have you published your tradesecrets
if they are encrypted?

IMHO yes. From my point of view tradesecret status is 
lost if there exists if and only if there exist
a function (F) that takes the publish work (W) and
transforms it to the tradesecret (T) and no function
(G) exist that transforms F to T. That is you can't
hide the tradesecret in transformation function otherwise
it would always be fullfilled.

Whether function F takes millions of years to execute or
not is irrelevant IMHO. Published is published.

In short:
Exists F so F(W) = T and not exists G so G(F) = T
 
> > The main problem is that some things that the LGPL and the 
> GPL tries to
> > do is either not well defined (mostly LGPL) or obviously 
> incompatible
> > with copyright law (mostly GPL).
> 
> amen. :)
>  
> > Copyright law wasn't designed with copyleft licenses in mind so some
> > thing that isn't very relevant there normal copyright is concerned
> > is largely undefined.
> > 
> > Specifically patches (read: modifications) isn't a normally 
> a problem
> > with normal copyrighted work since in order to for an end user to
> > benefit from the modifications he must legally buy of copy of the
> > orginal work and thus the orginal author will always be compensated.
> > 
> > With copylefted works, which the end user usually can get for free,
> > this breaks down...
> > 
> > This is a much larger problem that I think most people realize,
> > since I can see no sane solution to this problem other than the
> > obvious one. "If you did't ask for compensation for each copy,
> > tough luck, you had your chance, don't come crying because
> > you made a bad choice. After all it was your voluntary choice".
> 
> That's if the legal system is being nice.  
>  
> > Of course, if so, it would mean that all copyleft licenses were
> > essentially equal to the X11 license...
> 
> Another possible outcome might be that if you require any 
> compensation,
> you lose any indemnifcation (and this is not too far 
> fetched).  

But no copyleft that I know of require anything from the end user so
the end user can't claim any compensation based of this.

As for non end users, that is people that modify the works.
Well, since they modified the your work they are probably
on their own since any accident might be their fault.

No I don't think that they would be worse of concerning possible
liabillity.

> If this
> happens, it would be *infinitely* worse than X11, and you are 
> forbidden
> to change the license to boot (unless you add that "later" clause AND
> FSF gives up copyleft - two big Ifs).  Never underestimate the legal
> system.

What do you mean?

Obviously I can change the license to something less restrictive on my code.
If I said that all my code release to the Wine project was retroactively
public domain it would be instantly and for all times to come.

Sure it wouldn't protect me if the code contained some sort of virus or
timebomb or whatever but that have nothing to do with copyright.