GPLv3 2(d) (was Re: PHPNuke license)
David Turner <email@example.com> writes:
> Can we please, please, please start another thread to discuss this?!
>> that's enough reason for
>> me to stop releasing code under "version 2 or later" of the GNU GPL:
>> the persistent spectre that future versions will prohibit certain
>> sorts of functional modifications.
> That would be silly, since you could always fall back to v2. The only
> reason to fear v2 or later is that v3 could be too permissive, not too
No; if I release software under v2 or later, and a v3 with this clause
is released, I have a problem: somebody can take my work, make
modifications to it, and distribute it in such a way that I cannot use
it. Even if he gives me the source code, I can't make use of his
modifications without "upgrading" the licensing of my code to v3. If
I'm going to give people the freedom to take my code and make it
non-free, I might as well just put it under an MIT license.
>> Wouldn't a requirement that if you make the software available for use
>> to another party, you provide an offer of source to those users make
>> much more sense, and avoid entanglements with the function of the
> That would be impossible under US copyright law, where "use" isn't one
> of the 17 usc 106 exclusive rights, while modification is.
Public performance is restricted by copyright law; I'd certainly
consider an Apache web server to be a public performance of Apache.
In any case, there's another problem I allude to above but didn't
mention clearly: the existing requirements for source distribution are
very flexible. This proposed 2d imposes technical limitations on
Every time a license tries to use exact technical definitions, it ends
up breaking a few years later. I'm not nearly as worried about the
HTTP requirement as I am about the definition of "computer network".
Is my USB keyboard/mouse a network? How about my Bluetooth keyboard?
When I'm interacting through an anonymizing mix-net, there's a decent
chance I'm not online when the other side is. Are we interacting
through a network? What about a web client which responds to certain
server requests for its source: it may not have any way to hear a
request from the server, and if it's used as the skeleton for an
embedded control system, all that junk needs to go along with it.
I'd far prefer to see a GPLv3 grant and guarantee more freedom, not less:
* Remove technical requirements such as 2c, the object/executable
langauge in 3, the header/Makefile and OS exceptions in the later
section of 3.
* Remove the strange definition that a work containing nothing both
creative and derived from a GPL'd work be considered a derivative of
the GPL'd work: that is, remove the definition that linking is
modification. It's not in the license now, but clearly state that
if you incorporate nothing creative from the GPL's work, you are not
a derivative work.
* Add "public performance" to the scope clause in 0, permitting (for
example) me to give a lecture on the details of GNUtls, or to run a
web server which presents an interface to Perl.
 That is, modify and distribute it in non-free ways.
 Admittedly, an odd case.
Brian T. Sniffen firstname.lastname@example.org