[License-discuss] GPL linking exceptions

[Rick Moen]

Quoting Chris Travers (chris at metatrontech.com):

> In the GPL v2 it is less clear.  At its outer limits however, it can
> only permit things prohibited under copyright law.  Generally RMS
> seems to think this is not permissible, and most other people outside
> the FSF don't listen.

FSF folks occasionally spend time advocating behaviour they would like
to see occur when the actual question is whether a course of action
violates someone's copyright rights or not.  I can only suggest careful
and skeptical listening to hear _why_ they assert something is tortious.

I like your characterisation that GPVv2 (and, I would add, any licence)
can only permit things prohibited under copyright law -- concerning
an instance of someone's creative work (for which certain reserved rights
would otherwise default to the copyright owner and not be granted at all).
When a second party then creates a derivative of that work, adding
his/her own copyrighted elements, FSF are welcome to have an opinion
about what permmissions the second party is permitted to grant
concerning his/her property, but I can't recall hearing credible
reasoning supporting this opinions.  (OTOH, points they make about 
the second party lacking ownership over the reused work and thus not
having the right to grant additional permissions to _it_ are well
taken.)

Unfortunately, you appear to be advising Mr. Krause that adding a
'linking exception' would be tolerable not merely concerning Mr.
Krause's copyright interest but also would somehow fail to infringe the
borrowed work's copyright interest.  It seems to me that such can be
true only if _either_ the borrowed elements are ineligible for copyright
coverage _or_ aren't pervasive enough to trigger the applicable test for
'substantial similarity' (abstraction, filtration, comparison for
non-literal copying, probably 'total concept and feel' for literal
copying).  In any specific case, those are, as the saying goes, triable
questions of fact.  



> Additionally there is the question (much discussed on this list not
> too long ago) whether copyright law even allows restricting secondary
> markets for practical tools in this way.  Rick Moen and I strongly
> disagree on this issue.

I said nothing about 'practical tools'.  The usual points of contention
in the more interesting sorts of copyright infringement cases under USA 
law are whether copied elements (if they rise above the rather low bar
for originality) are functional or expressive, and whether an allegedly
expressive element is more properly characterised as an idea or method
and thus not eligible for copyright at all (albeit potentially
patentable).  

Caselaw in this area (involving literal copying) typically works out the
way Apple v. Microsoft did, where the Ninth Circuit flatly rejected
Apple's assertion of copyright protection over the way MacOS used icons,
program windowing, pull-down menus, and distinctive widgets, saying
Apple could not get 'patent-like protection for the idea of a GUI' --
the idea/expression divide from Baker v. Selden, Feist v. Rural, etc.  
You cannot copyright an idea; you cannot patent an expression.  This was
likewise the key point in Lotus v. Borland, where the court held that 
the program menus Lotus sued over embodied the _methods_ used by the 
program to operate, and therefore Quattro Pro's verbatim copies of 469
words in 50 Lotus 1-2-3 menus did not infringe.

> I think the clear pattern of cases including Oracle v. Google points
> towards a negative answer here.

I think you are drawing a rather sweeping and seemingly non-sequitur
conclusion.