🚀 go-pugleaf

In article <BURLEY.92Oct4011851@geech.gnu.ai.mit.edu> burley@geech.gnu.ai.mit.edu (Craig Burley) writes:
>In article <1992Oct3.221703.1496@fcom.cc.utah.edu> terry@cs.weber.edu (A Wizard of Earth C) writes:
>
>   o	Potential legal costs if UCB, CMU, or GNU get upset about you selling
>	   their code (GNU is real uptight about this, and CMU and UCB could
>	   easily get uptight if AT&T can prove licence violation, one of the
>	   charages against UCB).
>
>What evidence do you have that "GNU is real uptight about [someone else
>selling their code]?"

	If this were the issue I was trying to address, I would point out
that you have disallowed the best example as "not counting" below, and that
any of the weaker arguments I could give you would be likely to be dismissed
out of hand as well.

>It'd be kind of strange for them to be uptight about something they went to
>a fair amount of trouble to explicitly permit in the licenses for their
>software.  (I've seen a lot of free software that simply says "can't
>be distributed for commercial gain" -- the GPL has no such restriction.)
>
>If you don't have any actual, hard evidence that Project GNU has ever
>hassled somebody for _selling_ GNU code (as in charging for distributing
>GNU code and/or programs to customers), please just say so, and don't start
>a flame war.  A case where Project GNU (really the FSF) hassled someone
>because they wanted to distribute GNU-derived code with _different_
>(especially more restrictive) licensing agreements for a fee does not count,
>of course, but that should be obvious given the context.

	I was talking about what it would take to be able to break even
plus pay a small amount to the person doing it for "doing a distribution
right".  One of the issues is defense against legal hassles.

	Right now the biggest legal hassle facing a for-profit distribution
is the USL/BSDI suit... but even assuming that BSDI wins and there aren't
any barriers to distribution from that quarter doesn't mean that there won't
be problems otherwise.

	I was simply giving basic issues in a risk-assesment exercise to
see where we came out in dollar amount for a professionally managed "free"
distribution.  What raises the cost is not that the FSF has abused the
lack of clarity in the GNU public license, but that the lack of clarity
exists and must be weighted.  The statement "GNU is real uptight about this"
was prefaced by "Potential legal costs if ... GNU get[s] upset".  The
reference was thus modified to refer to their being uptight about the idea
of "selling their code", not uptight about the *use* of their code.

	If I am essentially "selling" a "free" distribution, it doesn't
matter if I call my activity "charging copying fees", "documentation and
support fees" or something else entirely; it is likely that some part of
my profit is attributable to the GNU code.  Article 2(b) of the GPL puts
any work containing it as a component in question if I bugfix the GNU
code so that it will work a particular way for 386BSD.

	The problem arises from the non-specific wording of the license and
the fact that I can not make a profit on the distribution of the code.  Does
my profit derive from some portion of the  code rather than the code as a
whole?  I can possibly make a case for this if the GNU code is partitioned
as a seperate element.  I can probably even hedge this by charging double
"copying costs" -- once for the binaries, once for the sources.  But I can
not make a profit... although this may violate 3(c).  This is risk reduction,
not risk elimination.

	Since the GNU software is not a seperate element in either 386BSD
or Linux, this put me in a difficult situation.  The license disallows me
making a profit, but I may charge a "reasonable copying cost", but my
idea of reasonable might be "Pays for the person copying plus the shipping
plus my time for the application of my management skills plus my flooring
plus my taxes, etc."... GNU hasn't really defined what they deem reasonable
to mean.  I could very well be considered a "profiteer" due to my definition
not matching an unpublished definition.  Article 1(last) implies that I
can only charge for the person doing the copying, and that any organizational
overhead can not be added to the cost.  This is basically an interdiction
of commercial copying, and requires a plausable "seperate element" argument,
per the paragraph above.  This is further muddied by the 2(d) in that I
may have costs which are not attributable to "physical copying" or "warranty
protection".  I might be able get around this by stating that I have to be
in business to offer warranty protection, but this forces the 2(b) bugfix
issue.

	So we are stuck on the horns of two dilemmas, as it were:

1.	When is my revenue in a for profit operation deriving from GNU
	code instead of other code (and more important, where does GNU
	believe this demarcation occurs)?  Is it when I make a profit
	off of software patches requiring GNU compilation at the user
	site, for instance?

2.	What is "reasonable" as it is used in the GPL?



	In my risk-assessment, I actually assigned this the lowest risk of
any potential legal action, if that will help salve the wound.  I would,
however, be an idiot if I didn't consider it at all; after all risk is
the relative *potential* that something will go wrong.

	The highest risk, of course, is that AT&T will simply drop it's
suit against BSDI.  That way, there is no binding case law preventing
AT&T from going after a smaller fish (potentially myself) to get their
binding case law.  Why would they do this?  Less legal difficulty in
getting what will amount to a pro-forma decision on the basis of the
small fish being unprepared to defend itself as rigorously as BSDI might.
My best defense would be to attempt to get BSDI involved again to try
and make myself look like a big fish.

	The next highest risk is that AT&T loses, but wins a pyhric
victory in their licensing issue against BSD.  BSD has two choices:
Buy an AT&T license (and contaminate themselves) so that BSDI and my
hypothetical company can continue commercial use of the code (thus
leaving the Berkeley license unattacked), or withdrawl their code from
commercial use (admitting a hole in the Berkeley license, but avoiding
the AT&T license issue).  This second has the advantage that it would
probably be offered by AT&T as an inexpensive compromise to show thier
"good faith" and still get what they want.  This would be a much better
deal from AT&T's point of view.  Rather than a single commercial license
fee, they forego payment to get a lock on their "competitors" technology
pipeline.

	There are many more gradations before the "GNU risk", so it's
pretty low on the totem pole.  But it *must* be considered.



					Terry Lambert
					terry@icarus.weber.edu
					terry_lambert@novell.com
---
Any opinions in this posting are my own and not those of my present
or previous employers.
-- 
-------------------------------------------------------------------------------
                                        "I have an 8 user poetic license" - me
 Get the 386bsd FAQ from agate.berkeley.edu:/pub/386BSD/386bsd-0.1/unofficial
-------------------------------------------------------------------------------