[lucy-dev] Only required notices in NOTICE

[Marvin Humphrey <ma...@rectangular.com>]

Greets,

I have committed the changes described in this email as repository revisions
r1208864 and r1208868.  If no one objects within the next couple of days, I
claim lazy consensus.

Marvin Humphrey

-----------------------

Even former ASF attorney and current ASF board member Larry Rosen isn't clear
on what goes in NOTICE:

    http://s.apache.org/MBM

    FWIW, I have never understood the distinction between the NOTICE file and
    the LICENSE file. I guess I assumed that someone else here did and that
    ASF had already explained it precisely somewhere. Based on this
    multi-year-long thread on JIRA, I now assume I'm not the only one
    confused. 

    Is there a link somewhere that explains what goes in NOTICE and what goes
    in LICENSE? 

The documentation was recently updated after a long debate on
general@incubator revealed the state of confusion:

    http://apache.org/legal/src-headers.html#notice
    http://apache.org/legal/resolved.html#required-third-party-notices

Based on the more minimalist requirements for NOTICE which have emerged from
the clarifications, we can remove a few entries from ours.

First, we can move the public domain information about the Lemon Parser
Generator and the US Constitution sample text out of NOTICE and into LICENSE.

  This product bundles the text of the Constitution of the United States of
  America, which is in the public domain.

  This product bundles the Lemon parser generator from the SQLite project, which
  has been released into the public domain by its authors as explained at
  <http://www.sqlite.org/copyright.html>.

Second, we can remove the Snowball copyright information, because the
inclusion of the Snowball's copyright notice and 3-clause BSD license in
LICENSE suffices.

  This software bundles code developed by the Snowball project at
  <http://snowball.tartarus.org>, Copyright (c) 2001, Dr Martin Porter and 
  Copyright (c) 2002, Richard Boulton.

Lastly, there is the notice for the modified Unicode Data files in utf8proc,
and perhaps soon in our own generated files:

  This product bundles derivatives of Unicode Data files subject to the 
  following notice: Copyright (c) 1991-2007 Unicode, Inc. All rights reserved.
  Distributed under the Terms of Use in http://www.unicode.org/copyright.html.

I originally believed that this notice was required in order to satisfy the
"associated documentation" part of the Unicode license:

  (b) both the above copyright notice(s) and this permission notice appear in
  associated documentation,

However, I no longer believe that it makes sense to include that snippet in
NOTICE.

  * If Lucy has no "associated documentation" (because the bundled docs are
    part of the software), we're fine.
  * If LICENSE is considered "associated documentation", we're fine.
  * If LICENSE is *not* part of the "associated documentation" but NOTICE
    *is*, then we have to include the full Unicode license in NOTICE as well
    as LICENSE.  That's silly.

My position is that that our inclusion of the complete Unicode license in
LICENSE suffices.

What we are left with in NOTICE after all that is the ASF notice, which is
unambiguously required per ASF policy:

    Apache Lucy
    Copyright 2010-2011 The Apache Software Foundation

    This product includes software developed at
    The Apache Software Foundation (http://www.apache.org/).