[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: using GPL api to be used in a properietary software

From: David Kastrup
Subject: Re: using GPL api to be used in a properietary software
Date: Mon, 14 Mar 2005 11:51:57 +0100
User-agent: Gnus/5.11 (Gnus v5.11) Emacs/22.0.50 (gnu/linux)

Stefaan A Eeckels <> writes:

> On Sun, 13 Mar 2005 18:59:23 +0100
> David Kastrup <> wrote:
>> Stefaan A Eeckels <> writes:
>> > Tell me to respect the wishes of the author, and I'm all with you,
>> > even if these wishes seem - at first sight - rather outlandish.  But
>> > this lunatic fight to get the scope of copyright extended, by
>> > exactly those people who originally wanted to abolish all forms of
>> > copyright, is one of the saddest quixotic battles I know of.
>> You don't get it.  The FSF is not fighting for the laws that give the
>> GPL teeth.  But while this insanity prevails, nothing is gained by
>> pretending it isn't.
> Have you stopped to think about the implications of having dynamic
> linking (where, remember, nothing more than a number of references
> to a library are contained in the compiled code) legally equated
> with producing a derivative work? It would be tantamount to
> declaring all source code derivative works of the OS.

Where the source code is useless without the system, this position is
tenable.  If you are writing for an API for which more than one
implementation exists, this is not as likely.  If we take the GPLed
readline library as an example, if I remember correctly, some party
implemented a non-GPLed readline clone that was source-compatible.
Once this clone was available, the GPL of the readline library was
effectively turned into the LGPL: it did no longer extend to source
code that could be compiled to use either library.

In a similar vein, if you are programming for published standard like
Posix, and possibly one that has several implementations, you can't be
held to be deriving from any particular implementation of the
standard.  Where you are concretely interfacing with special
Linux-only kernel functionality, things would get more muddy without
the kernel exception.

> What's the difference between a function call in source code, and
> its compiled counterpart?

Who claimed a difference here?

> OK, the GPL didn't take dynamic linking into account in its
> strategy, and the effect of GPL'ing libraries was less than
> expected. But then stubbornly pursuing a strategy that, when
> successful, would be an effective strengthening of the restrictions
> the copyright statutes already impose, shifting the balance even
> further towards the large corporates, can only lead to a Pyrrhic
> victory.

You are not getting it.  Really.  It is not the strategy of the FSF to
strengthen the copyright statutes, but merely don't use less than what
is claimed by other parties and established in the court already.  The
FSF would like nothing more than legal and unanimous precedent that
says that linking to unique interfaces does not constitute derivation
in any manner.  Whether this precendence comes about by a court case
against the GPL itself or any different licence, is a secondary

>> If you want no defense against people unilaterally taking your work
>> and turning it as proprietary as the laws allow, use the BSD
>> licences.
>> The explicit and expressed purpose of the GPL is to make the code
>> it covers not be subvertible in this manner.
> The use of a GPLed library doesn't subvert the code. It fails to
> extend the GPL to the program, but the whole take of the FSF on
> "user does the linking" is merely sour grapes (doesn't the GPL
> itself not say that it doesn't limit the user from using the
> program?).

The GPL covers copying and redistribution.  If no use can be made of
the product except by linking it to an FSFed library, then the
responsibility, of course, also rests with the distributor.  In a
similar vein, a weapon manufacturer can't circumvent weapon law by
shipping weapons and ammunition separately and claiming that it falls
entirely into the user's responsibility to combine two completely
harmless items into something covered by arms' laws.

> This isn't about Alexander. This is about risking to get judgements
> that will throttle any and all independent software developers even
> more effectively than the current hideous patent initiative of the
> Council and the European Commission.
> You wanna write an app for our OS? Ask our permission first. Thank
> you.

Good reason to switch to a free operating system, if it does not yet
suffice that MS reserves the right in its EULA to destroy your system
remotely in the interest of Digital Rights Management without being
held accountable for any damages.

David Kastrup, Kriemhildstr. 15, 44793 Bochum

reply via email to

[Prev in Thread] Current Thread [Next in Thread]