On Tue, 22 Sep 1998, Greg A. Woods wrote:

> [ On , September 21, 1998 at 23:00:40 (-0700), Chris G. Demetriou wrote: ]
> > Subject: Re: More licensing flames...
> >
> > woods@most.weird.com (Greg A. Woods) writes:
> > > The Sleepycat copyright license on their DB-2.x goes one step further
> > > and contains terms that some might describe as "viral" -- i.e. they
> > > attempt to force people who distribute software linked against DB-2.x to
> > > also be freely re-distributable in source form.  I.e. the DB-2.x
> > > copyright license is very much more like the GNU LGPL than the GPL.
> > 
> > Uh, do you have those reversed?
> 
> Nope.

Uhm, as I understand all of this, the sleepycat license is more like GPL
than LGPL.

> > The LGPL says, in a nutshell: you link bits against these LGPL'd
> > sources, you have to distribute the LGPL'd sources and at least
> > objects for the rest of the bits (not necessarily sources).
> > 
> > The GPL says, in a nutshell: you link bits against these GPL'd bits,
> > you have to give away sources for everything.
> 
> I take it you're referring to section 2 paragraph (b) of either version
> (or perhaps section 10 of the GPLv2 or section 8 of the GPLv1), and I
> think you're taking it out of context of the copyright license as a
> whole.

I thought these sections (as paraphrased above) are the whole point of
(L)GPL, and the reason we can't distribute kernels w/ them in it.

> Most specifically the GPL makes it quite explicit that it does not
> intend to restrict the rights of any other programmer.  Note the
> explanatory paragraphs include in section 2, and in particular this
> sentence (from GPLv2):
> 
> 	If identifiable sections of that work are not derived from the
> 	Program, and can be reasonably considered independent and
> 	separate works in themselves, then this License, and its terms,
> 	do not apply to those sections when you distribute them as
> 	separate works.
        ^^^^^^^^^^^^^^

I don't think that anyone has been arguing that linking w/ (L)GPL'd stuff
means you loose the copyright on your code. If they were, I missed it. 

The thing though is that (L)GPL has a restrictive LICENSE on how you can
distribute such code.

> Besides, it refers to using "parts" of the Program, and implies
> derrivation.  Most times a library, esp. a specific library such as DB,
> is used in an application the entire content of the "Program" is
> included with the distribution of the application that uses it.  The GPL
> section 2 attempts to "control the distribution of derivative or
> collective works based on the Program".  Including the "Program" as a
> library which a larger proprietary application uses does not form a work
> "based on the Program", but rather one based on the application owned
> entirely by proprietary software vendor.  I.e. the emphasis is extremely
> important.

Actually, sticking the program inside another program would be covered by
section 2 as I understand it. That's what GPL is trying to prevent. You're
right that the proprietary software house would still own all the other
parts. That's not the problem, though. They couldn't legally redistribute
the program if it included parts which were from GPL'd code UNLESS they
made the source for the whole thing available. 

_THAT'S_ the problem.

> One thing that's very important to keep in mind is how the copyright law
> treats "collections" of works.  I've not studied the US specific law in
> any detail, but I have studied the Canadian Copyright Act and much of
> the interpretation that's been applied to it in this country.  It is my
> un-official expectation that any two separately copright licensed works
> which are combined into one "binary program" will be treated as a
> "collection" and that a separate (and possibly very restrictive)
> copyright license (eg. "All rights reserved") may be used to protect the
> collection as a whole, but that the copyrights of the individual works
> contained in this work will still apply to them as a portion of the
> collection.  It is *impossible* for any copyright license to affect
> other individual works that are included in the collection and a
> copyright license may only affect the copyright of the collection as a
> whole to the extent that this affect would only apply to the individual
> work in the collection to which it applies.

It is though perfectly possible for the license to dictate how the
collection is distributed, though. 

> For example consider what might happen if a poem that's got a copyright
> similar to the GNU GPL is incorporated into a book or magazine that is
> covered by an "All rights reserved" copyright.  This book or magazine
> would not be "based" on the poem.  The author of the book or magazine is
> free to sell this collective work for a profit.  However I or anyone
> else can freely distribute copies of the pages on which that poem
> appears, but we cannot distribute copies of the entire book or magazine,
> and we cannot necessarily distribute copies of any other poems or
> articles in the book regardless of whether those other parts were
> written by the author of the collection or by third party authors.  The
> authors of any other works that are included in this book or magazine
> may still be free to distribute copies of the pages on which their works
> appear (though this will depend on the exact terms of the one-time
> license they granted to the author of the collection).  In fact if I'm
> remembering right we can even copy any "small" preamble that the author
> of the collection may have included at the beginning of the poem because
> that would be considered "fair use" (so long as I play by the "fair use"
> rules, of course, and acknowledge the source of the preamble).

Actually, I think you couldn't copy THE PAGES from the book, but you could
copy the poem. I.E. you could type it into another book if you gave
credit.

I say this as it was brought to my attention that a number of legal
journals (reporting public court decisions, etc) copyright the page
numbers in their documents. They can't copyright the decision (it's
public), but they can do something to protect the effort they took to
making the book.

> Remember the intent of the entire GPL is to protect the "Program" and to
> ensure that it remains free and that every identifiable derrivative of
> it also remains free.  As I mentioned above the GPL makes quite explicit
> that it does not intend to restrict the copyright rights of any other
> programmer who might also be a "user" of free software.

Agreed. 

> However since the normal GPL won't permit the user of a proprietary
> application to modify the free library and then re-link it with the
> proprietary code, the FSF decided to come up with the LGPL which would
> force proprietary software vendors to distribute their software in a
> form that would permit re-linking with any free libraries they use.

?? Wasn't the reason for LGPL so that proprietary programs COULD use free
libraries? For instance, so that Matlab or Netscape could be legally
ported to Linux? If it weren't for LGPL, proprietary code couldn't TOUCH
GPL'd libraries.

I thought the point of LGPL was to draw a line and say that the license
restriction of the GPL only went so far. Kinda compartmentalizing the
program into a free and a proprietary part.

> >From a service and support perspective the LGPL is enormously more
> restrictive because it raises the number of potential problems that the
> end user can encounter by many many orders of magnitude.  I.e. the
> degree of "freedom" which the LGPL gives to end users is potentially
> totally stifling to a proprietary software vendor.
> 
> The FSF puts the GNU LGPL on libraries because they feel that the user
> should be able to modify the library even when doing so may totally
> compromise the integrity of the application as a whole.  (Eg. the author
> of the application makes explicit compensation for a bug in the library;
> compenstation that would cause the application to fail should the bug
> actually be fixed.)

I thnk then that the application author just needs to say that if you
change the code, you're on your own.

Besides making commercial software possible for Linux, I think the LGPL's
useful in that if, for instance, a program is dynamically loading little
LGPL'd modules, I can extend the functionality of these modules. Say I can
add an operating mode which wasn't avalable before. :-)

I'll admit it's a contrived example. :-)

> (FYI, I personally don't like the LGPL, but I do like the GPL.  I don't
> see any valid reason to give users of software the degree of freedom
> that the LGPL affords them.  Most "users" are incapable of making proper
> and safe use of this freedom.  Maybe that's my Canadian way of thinking
> showing through....)

I agree that for the case of a library linked into a program on a
proprietary/*BSD OS, it's dangerous. But since LGPL also permits a lot of
the big commercial packages I use at work to make versions for Linux
(which unfortunatly is much more likely than making versions for NetBSD),
I think it's a good thing.

Take care,

Bill