----- Original Message -----
From: "Alan Hodgson" <ahodgson(_at_)simkin(_dot_)ca>
To: <spf-discuss(_at_)v2(_dot_)listbox(_dot_)com>
Sent: Wednesday, August 25, 2004 1:19 PM
Subject: Re: [spf-discuss] Re: [Fwd: DEPLOY: Microsoft Patent license
unworkable with GPLed MTAs]
On Wed, Aug 25, 2004 at 09:51:47AM -0400, Holm, Mark wrote:
I am not saying the MS license should be accepted as part of an RFC
package
without careful examination and consideration. I may well be missing a
critical aspect that causes trouble. But a lot of the things people
have
been writing seem to me to be the result of sloppy reading or thinking
about
the licenses. Take your time. Read it carefully. Read the GPL
carefully.
OK, it's 2006. You download the Fedora Core 4 DVD images. On those DVD's
are,
maybe 5, maybe 20 different packages that have to make use of Sender ID to
function on the Internet of 2006. These packages include source code.
RedHat won't include that kind of software at all in Fedora. They put that
sort of licensed software in the Enterprise releases.
You wish to burn copies of those DVD's for your fellow LUG members. But,
wait, now you have to sign a license agreement with Microsoft before you
can
do so. RedHat had to sign a license agreement before distributing the
packages to you. Every Fedora mirror operator had to sign a license
agreement with Microsoft to host the packages. Every other distribution
has the same situation.
This sort of fun and games, although in the copyright venue not patent, is
what just killed the XFree86 distribution and caused major vendors to switch
to using xorg instead. If this sort of patent licensing is mandated in the
standard, expect a software fork to provide Classic SPF in one package and
SenderID in an entirely distinct package so that people can gain the
benefits of SPF without having to cope with the licensing difficulties.
Now, fast-forward another few years. Since the IETF allowed this to
happen
once, they will allow it again, and again. In 10 years you may have to
sign
50 or 100 different license agreements in order to use the code
required to function on the Internet.
That's exactly the sort of layered intellectual property agreement craziness
that used to occur in BSD and which caused such turmoil in XFree86 more
recently.
Can you not see the problem here? Do you really think the open source
community will allow this to come to pass?
Expect a fork.