The entire article and more or less all the high-mod'ed postings
seems to be based on a weird premise: That reproducing and imposing
an extra set of license terms is not a "further restriction" (clause
6).  Reproducing an extra set of license terms is a "further
restriction".

From my experience working with *BSD and seeing proprietary derivates
done, this can actually be a major hassle.  It is a lot of work to
just extract the individual licenses for reproduction, and it is
even more work to run those licenses through the lawyers to check
that they are OK etc.

In practice, this may easily run in the thousands of dollars.
Definately noticable, and having to do this definately counts as a
"further restriction."

I've seen the standard defense of the people that like to grap code
under other licenses and slam it together with GPLed code: This is
not a restriction due to section 1 of the GPL (... publish on each
copy an appropriate copyright notice and disclaimer of warranty
...)  However, this statement says *an appropriate copyright*, not
*a specific copyright and disclaimer as limited by a separate extra
license*, which becomes seriously noticable when there are many of
these.


The only reason I can see that the FSF does not count it as one is
that they actually do not have the experience working with these
licenses in a carefully controlled manner.  This experience generally
comes from being somehow involved with proprietary derivates of
free software - e.g, through being involved on the development side
of *BSD (where we see proprietary derivates as a GOOD thing, as we
get a lot of good code back from proprietary derivates.  We actually
have to do work to keep the number of variants of the BSD/MIT style
licenses down, because it becomes onerous for derivates to reproduce
them all.)

Also, the FSF is NOT the supreme authority on what is compatible
with the GPL and not.  The courts are.