Mailing-List: contact legal-discuss-help@apache.org; run by ezmlm
Precedence: bulk
Received-SPF: pass (hermes.apache.org: local policy)
Errors-To: <wrowe@rowe-clan.net>
Message-Id: <6.2.1.2.2.20050329232236.052d9d60@pop3.rowe-clan.net>
Date: Tue, 29 Mar 2005 23:29:00 -0600
To: Jeffrey Thompson <jthom@us.ibm.com>
From: "William A. Rowe, Jr." <wrowe@rowe-clan.net>
Subject: RE: Apache License : grant of patent license and derivative
  works
Cc: <lrosen@rosenlaw.com>, legal-discuss@apache.org
In-Reply-To: <OF7BD1483E.3AE464F5-ON05256FD4.000FD3F8-05256FD4.00120735@
 us.ibm.com>
References: <20050329191104.GA22021@mail26f.sbc-webhosting.com>
 <OF7BD1483E.3AE464F5-ON05256FD4.000FD3F8-05256FD4.00120735@us.ibm.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"

At 09:16 PM 3/29/2005, Jeffrey Thompson wrote:

>I'm normally quite a pessimistic guy, but I wouldn't be as pessimistic as that either.  The patent license applies to combinations of your Contribution combined with the Work.  "Work" doesn't mean only that specific version of the code that existed at the precise time that you made your Contribution.  If it did, then the patent commitment would be meaningless because changes are made every day.  Work means the code that is licensed under the Apache license to which the code was Contributed, or the "Project".  So, if you contribute to V1, the V1 is the Work and your patent license applies to your Contribution combined into V1.  When V2 comes out, V2 is also the Work and the license applies there as well. 

I agree, but consider the following (real) scenario.  Patents don't
apply, but easily could in a similar case.

Apache HTTP Server uses a patent (not really) that describes a method
to close a socket in a graceful manner, by waiting a certain number
of seconds before performing a 'hard close' irrespective of if the
client was waiting, or went away.  We (really) call the concept 
lingering close.

Apache HTTP Server 2.0 is released, and does the same thing for http:
sockets again as it did in it's 1.3 version.

Someone comes along and writes an ftp: module for Apache.  It uses the
lingering close feature.  Unfortunately, the grantor of this (non)patent
never intended to give it to the ASF for any purpose other than http:
and did not license that application of the method.

So we can easily be tripped up.  That said, the case doesn't vary that
much from the case where someone patented the lingering close method
unbeknownst to the ASF, and serves us a C&D to quit doing it.  I would
expect the grantor who willing offered us the lingering close method
to come back and state 'hey, that was for http: - what have you done?
Please quit it!'  And so, we would.

Now, I think a better solution is for the CCLA/ICLA to force the
individual to not only disclose the patent granted but it's scope
of fair use, so there is no ambiguity, and document same in NOTICE
for the particular project.

But now I'll let the real legal folks disagree with me :)

Bill 


---------------------------------------------------------------------
DISCLAIMER: Discussions on this list are informational and educational
only, are not privileged and do not constitute legal advice.
---------------------------------------------------------------------
To unsubscribe, e-mail: legal-discuss-unsubscribe@apache.org
For additional commands, e-mail: legal-discuss-help@apache.org