Re: OSD && DFSG convergence
On Tue, Mar 04, 2003 at 06:00:23PM -0500, David Turner wrote:
> > Your stretch relies upon a single act being both an act of distributing the
> > *modified* program and of invoking it interactively.
> I see no reason this can't be true of some programs. However, I do not
> *rely* on this (see below).
> > Here's why that must be the case:
> > * (2)(c) applies only when you're distributing a modified copy of
> > the program
> No, it applies when you modify the program. Please read the header for
> (2) again. It says, "You may modify ... the program ... provided that
> you also meet all of these conditions." Some would claim that I skipped
The elipses hide what is the crux of my argument. Emphasis added:
2. You may modify your copy or copies of the Program or any portion
of it, thus forming a work based on the Program, and copy and
DISTRIBUTE SUCH MODIFICATIONS or work under the terms of Section 1
above, provided that you also meet all of these conditions:
Now, I derive a different interpretation from this than you do. I
interpret this as:
* Any person with GPL code on his disk has extensive abilities to modify it
in almost any way desired, so long as it is not given to anyone. If I
want to remove the copyright statement from my private copy of bc, fine;
I just can't give it to anyone else like that.
* If the modified code is distributed to anyone, then it must adhere to the
conditions in both section 1 and section 2.
* Sections 1 and 2a through 2c do not apply if no distribution is taking
> a scope closing there, and that (a), (b), and (c) don't apply unless
> there's distribution. If that were so, why would (b) specifically
> confine itself to distributed versions?
I think it's just for clarity. Somebody reading the license seeing "You
must cause any work that in whole or in part contains or is derived from the
Program or any part thereof to be licensed as a whole at no charge to all
third parties..." might interpret this to mean that they are forced to
distribute and license changes that they make solely in-house.
The FSF GPL FAQ, in fact, mentions this itself:
> > * The "modified program" distribution could be modified without ever
> > having any human interaction! Can you imagine the dastardly
> > possibilities if you count as modification and distribution under
> > (2) acts that nobody even intended to perform or did perform?
> > I repeat: this "modified" program need NOT be modified by any
> > human. I maintain that this is not really a modification at all.
> things to happen. They may do so indirectly -- for instance, I could
> build a Rube Goldberg machine to modify some software. But they are the
But your intent is stil to do that, as with your autonaptster example. In
this case, nobody's intent is to do that.
> > * Even if it is modification, the license states "YOU may" (emphasis mine).
> > It does not give a computer process the rights to modify the program.
> > Are you asserting the absurd condition that a PHPNuke process is
> > violating its own license by modifying and distributing itself
> > without the user's knowledge?
> It's clear that the *person* causing it to be modified is whomever
> installed it.
I think that's a stretch. Let's say you make a plain vanilla installation
of PHPNuke. You change nothing from the defaults. You remove no copyright
If you assert that the output from PHPNuke constitutes distribution of
PHPNuke, you have to agree that the output streem is a modified version,
since it will not be containing all of PHPNuke -- just parts, intermixed
with other data. But you, as the installer, made no modifications to
Who then made the modifications?
> It would certainly depend on how the viewing happened. Would you say
> that "viewing a piece of artwork" could not possibly constitute
> "distributing a modified copy of the artwork"? If so, why hasn't
> somebody napsterized the contents of every museum yet?
If I am sitting at my desk and I have a Picasso here, I can view it. That's
one act. If I give it to someone else, that's distributing it.
If I view a GIF of the artwork online, somebody *else* distributed it to me
first, then I got to view it. That other person did not both view and
distribute it. The other person only distributed it; I viewed it.
Now this analogy is a little shaky because the nature of software is quite
different from that of artwork, and you could rightly poke holes in the
above analogy, but I don't think that hurts the base case of software.
> > The HTML code,
> > inclusive of embedded ECMAScript, is nothing but user interface
> > controls.
> This is true of many GUIs, especially those which are front-ends to
> command-line programs. Your narrow view of copyright, while perhaps
> defensible from a political perspective, has little match with the
> reality of case law.
If I fire up a web browser and go to www.phpnuke.org, what do I see?
some buttons and links.
Now if I fire up lynx -source on the same page, I get an entirely different
I maintain that these are separate actions.
The first could be interpreted as an attempt to execute an interactive
The second could be interpreted as an attempt to download the source to an
But by doing the second, you're not attempting to interact with it; and by
doing the first, you're not attempting to save the code. In fact, in the
first instance, the code would just be a transient cache in the way that you
indicated was exempt from copyright restrictions.
> > * Even if you do consider this to be distributing a modified copy of the
> > source code, I advance that the act of "downloading a modified copy of
> > the source code" is fundamentally different from
> > starting and running for "interactive use in the most ordinary way" and
> > thus is not subject to (2)(c).
> In some cases, maybe. But in this case, I don't know what else IUITMOW
> could be.
Perhaps there is none in this case.