Dispelling BSD License Misconceptions (fwd)

[Brian Behlendorf]

This may be off-topic in so far as license-discuss is concerned, but since 
the quality of our analysis is important, I'd like to understand why 
reasonable minds may differ to such a degree that the following might be 
true.  Especially since IANAL and I'd like to think this was something a 
lay person could navigate...

In particular, I think his claim in his paper, section 3.3:

   It is likely that the reasonable person would read the license and think
   that the licensor intended that the warranty disclaimer was to run with
   the redistribution without qualification.

and then the immediate leap in 3.4 to:

   On this analysis, the warranty disclaimer travels with the distribution
   and the distributor has no ability to qualify it.

is where the author goes astray, and with it the rest of the argument. 
The need to "retain" the original license in any redistribution does not 
imply that any redistribution must be under that original license.  It is 
included to communicate to the recipient that some subset of the code, 
perhaps even all of it, was originally under that older license.  It 
doesn't preclude the possibility that there is a newer license, perhaps 
even to cover newer modifications or newer code.  Nor would a newer 
license prevent a recipient from exercising their rights around the older 
code under the older license.

That suggests that no, the warranty disclaimer doesn't automatically apply 
to the redistributor.  Most redistributors who care about this issue have 
probably separately disclaimed their liability for any third-party 
software they are redistributing, and separately for any of their own 
code. Another angle to this is that a redistributor could establish a 
code-insurance business, whereby they sell policies that assume liability 
for someone else's code, because they've vetted the risk against the price 
they think people will pay for such a guarantee.

At the same time, I have sympathy for:

   What the court is looking to determine is what the reasonable person
   (ie an idealized and dispassionate citizen who is called on to assess
   the scope of the license) would make of the words.

Would a reasonable person who found the BSD license in a bundle 
of code assume that the notice applied to the whole thing?  Is that what 
the original Berkeley Software Distribution copyright holders intended by 
the word "retain"?

He continues the mistake of confusing sublicensing with relicensing, it 
seems, by saying in 3.5:

   Moreover, if anyone could license the code under different terms, then
   any person redistributing the unmodified code could remove the warranty
   disclaimer and other terms, removing the protections expressly included
   by the copyright holder.

It seems to me that even if the license were modified, the protections 
would continue to exist for the copyright holder - if a redistributor 
removed the disclaimer and a recipient decided they needed relief, they'd 
seek that from the redistributor, and both the recipient and the 
redistributor would be unable to seek that from the copyright holder due 
to the original disclaimer from holder to redistributor.

I feel the author takes a further unsubstantiated leap when he says in 
4.3:

   However, the original licensor has mandated specific wording to be
   included which purports to apply to the code as a whole (ie as
   modified).

I don't read BSD that way, and I don't believe most others do either. 
Someone reading it for the very first time, without being introduced with 
a description of what the BSD license is supposed to mean, though, might 
be forgiven for thinking this.

Is this a case, though, where 14 years of reading a particular body of 
text and being told what it's supposed to mean can cloud one's sense of 
what it actually says?

 	Brian

---------- Forwarded message ----------
Link: http://slashdot.org/article.pl?sid=07/01/15/1757235
Posted by: ScuttleMonkey, on 2007-01-15 19:03:00

    [1]AlanS2002 writes "Groklaw is hosting an article by Brendan Scott
    which looks at the [2]misconceptions surrounding the BSD license. From
    the article: 'We observe that there exists a broad misconception that
    the BSD permits the licensing of BSD code and modifications of BSD
    code under closed source licenses. In this paper we put forward an
    argument to the effect that the terms of the BSD require BSD code and
    modifications to BSD code to be licensed under the terms of the BSD
    license. We look at some possible consequences and observe that this
    licensing requirement could have serious impacts on the unwary.'"

References

    1. http://alansplayground.spaces.live.com/
    2. http://www.groklaw.net/article.php?story=20070114093427179