BSD legal question

[Joel]

IANAL, etc.

> > But if we're going into this anyway, here's an interesting
> > point that people
> > tend to dabble about: One common misperception even in the
> > *BSD world IMHO is
> > that if you use and alter GPL code you have to release your
> > work under GPL
> > also. I don't think that's true.
> 
> Wrong, that IS true, here's the section of the GPL that dictates
> this:
> 
> "...But when you distribute the same sections as part of a whole
> which is a work based on the Program, the distribution of the whole
> must be on the terms of this License..."
> 
> http://www.fsf.org/licensing/licenses/gpl.txt

Not to forget that use, deployment, publishing, and distributing are
not all the same things. It is possible to use software without
publishing or distributing it.

One grey area is private/internal publishing, ergo, when a particular
piece of software is deployed within a company, some will argue that it
has been published internally, and must therefore be opened to all
comers. (And therefore if you do that and don't want to open your stuff,
you have to buy the non-free license.) Others will argue with that point
of view. 

Generally, since it is the author who selects the license, if the author
has made public a specific interpretation for his work, his
interpretation would likely get more consideration in court than the
user's interpretation, as long as it does not conflict with the rest of
the license or with established law.

This would be less of a grey area if 2.b were a little more specific
about all third parties. I think the legal definition of third party
would exclude people who were not party to the distribution or
publication, and that could override the issue of whether internal
publication invoked the license to a certain extent, assuming that
making the source available company wide were a feasible option for a
company. But, as I say, IANAL.

>  All that's required is that
> > you provide the
> > sourcecode to your changes (pedantically: only if asked for).

You do have to make your modifications available under the GPL if you
publish or distribute. You are not allowed to wait until the
modifications are requested. You do not have to force people to download
the sources, but you must make them available.

> > If you do that
> > under whichever license that pleases you, you're OK.
> 
> Not true.  The ONLY thing you could do is, for example,
> distribute a set of patches to a GPL program under the BSD
> copyright - but as such patches would be useless without the
> GPLed program,

You don't know that. In most cases they are probably not directly useful,
of course.

> as soon as the person that wants to use those
> patches applied them to the GPL program, the result would
> immediately become GPL. 

I would rather describe that as the GPL taking effect over the combined
or derived work. The GPL license on the combined work would not alter
the license on the separate work.

Of course, there is the "reasonably be considered independent and
separate" phraseology, which means that, if you wanted to assert a
different license on the patches themselves, you would do well for
asserting your basis for defining the patches as a separate work in the
same place you declare the license. I would imagine that most courts
would recognize an author's desire to keep the license separate as a
reasonable basis, even for a one-line patch, because the original author
of the patches also has implicit copyright on his or her work.

The counter argument would be that the one-line patch is meaningless out
of context and therefore must be a derived work. But I think US courts
would allow the author to assert a meaning. 

Of course, if the patch consists of just changing a single variable in
an expression, then the original source is still present, so the
single-line patch is in and of itself derivative.

> And of course, you could only
> distribute just those patches alone - not also distribute
> the GPL program.

You could do both for non-derivative modifications, but the distribution
of the combined/derived work would necessarily be GPL, where the
distribution of the patches alone could be the license of your choice.

> > By extension I consider the LGPL for things like libraries and
> > toolkits
> > effectively as BSD or MIT. Otherwise you might as well GPL all
> > content you
> > see with mplayer (which is legally impossible to begin with
> > because the
> > content was never theirs, same with original source code that
> > merely uses a
> > library or a toolkit or some other API that may be GPL'ed).

I could not parse this. However,

> That is a different deal - and one that hasn't yet been tested in
> the courts.  And the FSF has worked vigorously to keep such
> court cases as have arisen, OUT of the courts, and settled out
> of court.  The problem is such:
> 
> Suppose I distribute a library that is under my own copyright,
> yet carries a BSD-like license.
> 
> Suppose you then come along and take my library, and a GPLed
> library, link both of them together into a new program of yours.
> 
> The FSF says that the entire code now becomes GPL.

No, only the original author of a particular work can change the license
of the work. See the last paragraph of section seven where it attempts
to clarify that the GPL does not attempt to subvert the law that way.

However, most BSD-like licenses do not limit combining with GPLed code.
So the GPL would apply to the combination. But it the entire code does
not become licensed under GPL. The license information must be kept
intact in the combined work, and the part which was originally under the
BSD-like license could afterwards be separated again from the combined
work, and the original license would still apply.

> The problem here is that since you never owned copyright on
> my library, you do not have legal rights to modify the copyright
> and license on it.  Thus, you cannot legally apply GPL to it.
> Nor can the FSF or anyone else apply GPL to it.
> 
> Naturally, the parts of the program you are distributing that
> YOU wrote are under GPL.  But, suppose another guy comes along,
> takes your program, and pulls my library out of it and uses it
> for his program.  According to the GPL, if he does this that
> library is GPLed now, and his program must be GPL. According
> to the law, though, it's not GPL, it's mine.

From seven: "... a licensee cannot impose that choice."

GPL would only apply to the combined work. If it could not apply to the
combined work, the GPL does not give license to make the combined work.
And it does not give license to combing code under GPL with code under
incompatible licenses.

> And even more fuzzy is what happens if you put a dual-license
> on your resultant program which reads something like:
> 
> "This software is GPL if you use it as a whole, according to
> the terms of the GPL, but the parts of it that are under my
> copyright may also be used separately from the other parts that
> are owned by someone else, under a non-GPL.
> 
> According to the FSF's various FAQ's and such, you aren't supposed
> to be able to dual-license stuff like this - in short, once you
> distribute your program as a whole, it's all GPL.

That's sure not the way I read the stuff on their site.

If they say that, then why do they have that long list of compatible
and incompatible licenses? Why all the essays explaining the
incompatibities? Why all the warnings not to use code that has been
published under incompatible licenses?

> But once
> again this runs afoul of copyright law - because the "whole"
> distribution they are talking about in the GPL can be considered
> to be 2 distributions actually - one is the "whole" and the
> other distribution is a distribution of two separate parts - your
> part and the GPL part. 

In light of section seven of the license, how do you get this? Would you
mind quoting the pages that you are claiming say this kind of thing?

> That is, as long as you have coded things
> in separate modules, and licensed it properly, someone can come
> along and take just your modules out of the "whole" and use them.
> 
> Now you see one of the legal achillies heel's to the GPL.  That
> is why they wrote the LGPL frankly, not because of the public
> bullshit reason they state.  It is because code that is only
> libraries, not a separate program, invariably gets linked with
> other people's stuff, and your just asking for trouble if you
> try to viralize someone else's stuff.

If that were true, then why would Richard Stallman be encouraging people
not to use the LGPL if they can avoid it?

> This problem arises because the GPL is a LICENSE, whereas the
> BSD is a COPYRIGHT

That makes no sense at all. Copyright is copyright. If you want to talk
about this COPYRIGHT and that COPYRIGHT, you would either be talking
about the copyright laws of one country vs. those of another, or you
would be talking about specific rights defined under the laws of this
country and that country, or, most likely, you would be talking about
the copyright claimed by this person and the copyright claimed by that
person.

> that has come to be a generic term applied to
> any kind of software copyright license that reads like the BSD.  Strictly
> speaking, if I put:
> 
> "This software copyright Ted Mittelstaedt, Redistribution and use
> in source and binary forms, with or without modification, are permitted
> provided that the following conditions are met, blah blah blah..."
> 
> it is NOT a BSD license, it is a Ted Mittelstaedt license that
> reads like a BSD license, and so people refer to it as a BSD
> license.
> 
> People think that the BSD license works the same as the GPL license
> but they are two different animals.
> 
> In short, if I write code and retain copyright, even if I apply
> a license to it that is verbatim what BSD says, it's not a true
> BSD license.  It's a BSD-like license.  To get a true BSD license
> you have to assign copyright to the University of California,
> Berkeley.
> 
> > In any event, except when you use public domain, any copyright
> > declarations
> > and terms of acknowledgement stay valid by automagic copyright
> > law being
> > applied. IIUC the problem with public domain is that although
> > you discard
> > copyrights then you can still be held liable, but I'm not sure
> > about that.
> 
> The problem with public domain isn't that.  For starters you cannot
> discard copyright, there is always copyright.  What true public
> domain is, is copyrighted stuff where the copyrights have expired,
> or stuff that isn't legally able to be copyrighted,  or stuff
> that you have applied your copyright, then written in a license
> clause which states in effect that anyone has permission to
> re-copyright your stuff with their copyright.
> 
> That is the problem with it - because true public domain doesen't
> really exist in nature unless the stuff is uncopyrightable.
> (ie: news articles, jokes, etc.)
> 
> For example, consider The Holy Bible.  Copyrights on that have,
> of course, long since expired.  But, all a publisher has to do
> is rewrite it and bang - copyright instantly applies.
> 
> Of course, nothing stops another publisher from simply going back
> to the original Hebrew texts and translating them, or going back
> to the KJV (which the copyrights have expired on) and publishing
> that.  There is a rather amusing and interesting article on this
> here:
> 
> http://www.chick.com/information/bibleversions/articles/copyright.asp
> 
> Ted


--
Joel Rees   <rees at ddcom.co.jp>
digitcom, inc.   株式会社デジコム
Kobe, Japan   +81-78-672-8800
** <http://www.ddcom.co.jp> **