wxPython Licence vs GPL

Terry Hancock hancock at anansispaceworks.com
Wed Nov 23 14:27:12 EST 2005


On Wed, 23 Nov 2005 16:33:24 +0000
Steve Holden <steve at holdenweb.com> wrote:
> Whether or not some fragments of code remain unchanged at
> the end of  your project, if you start out with a piece of
> source code lifted from  wxPython then what you have
> created is definitely a "derivative work"  and, as such,
> you must take into account the wxPython license in your 
> licensing of the derivative work.

This is certainly a very pervasive claim, but I really don't
think it's true.  Copyright law (at least in the US), says
nothing about how a work was created -- only what the end
product is.  Copyright, after all, protects "expressions",
not "ideas".

What it means to "derive" a work from another is also
extraordinarily vague in fact -- I gather there has been
some definition through precedent, and I suspect it is the
use of provenance as *evidence* of derivation that promotes
this myth.  But the mere fact that you started with someone
else's work does not make it derivative.

Consider for example that there are fine art collages made
by cutting out magazine photos and assembling them -- in
most cases, the work is regarded as a new work. 
Furthermore, it is generally true that a work may quote
another work. But the interpretation is very vague and
highly inconsistent from medium to medium.

For example, it used to be that the same argument protected
sampled music and TV (and it is still common to find works
created under that assumption), but I believe that more
recent cases have attacked this and won.

Scholarly works quote copyrighted sources routinely, and in
general,  free speech must surely require this ability. In
the Sony DRM case, though, would this make quoting a GPL'd
piece of code in order to recognize it as a signature legal?
(I know that allegations have been made that they do more
than that, making the issue academic, but it's still an
interesting question).

The vagueness has been overly exploited and has effectively
allowed copyright owners to claim that "derivative" means
whatever they want it to mean -- and on that basis, maybe
the myth is true as a practical matter. But that isn't what
the law says. What the law actually says is that it has to
be "twenty percent different" -- but this is "in whole or in
part" so the implication is that it somehow has to be 20%
different *throughout the work*.

As regards software, there have been precedents to support
the claim that mere cosmetic alteration does not obviate the
copyright holder's claim, so the 20% change must also be
substantive, not just cosmetic.

But the fact by itself that you started with the original
work in your editor does not, by the letter of the law, make
it a derivative work.

The "clean-room implementation" concept is based on
alleviating a fear of litigation -- it forms the basis of a
defense. It is *not* a legal requirement, AFAICT.

Of course, IANAL. ;-)

Cheers,
Terry

-- 
Terry Hancock (hancock at AnansiSpaceworks.com)
Anansi Spaceworks http://www.AnansiSpaceworks.com




More information about the Python-list mailing list