Wow, this discussion is getting out of hand. It is not a technical
issue and thus isn't going to get resolved by throwing paint cans
at the shed. For years, the ASF had been following the examples
commonly seen in commercial software products of placing a general
copyright header all over the place to indicate the collective work.
This was fine when we started because we used to say
Copyright YYYY The Apache Group.
which, as a semiformal group of individuals holding joint copyright,
was both legal and correct. However, incorporating the ASF created
an entity that only held licenses (CLAs and license grants) to the
individual contributions. Thus, our old practice of copyright notices
should have changed accordingly. Unfortunately (or fortunately),
I am not a lawyer and did not know the finer details of US law
regarding copyright notices, and never had a reason to discuss it
with lawyers until last year.
I requested a related opinion from our ASF lawyers and, as part of a
lengthy discussion, was informed that we can't place a notice on
anything that the ASF does not own copyright (not just a license,
but ownership in the corporate name). So, then we tried adding the
"or its licensors, as applicable" suffix to the notice, which
satisfied 2 out of 3 lawyers (resulting in another round of finger
wrestling, and eventual agreement by all that it was bogus).
It was finally suggested that the Berne convention (and our liberal
license) made the notice unnecessary, and I came up with a minimal
replacement text for the existing header that simply informs
recipients of the licensor and terms.
A policy proposal was drafted by Cliff Schmidt, based on my proposed
header changes, and reviewed again (last October). That policy was
placed on the queue of things Cliff was going to present to the board.
Unfortunately, Cliff's free time got hit by the truck that we all
know of as the third-party licensing issue.
None of that, however, changes the fact that I (as an ASF officer)
asked our lawyer for a legal opinion and received an answer to the
effect of "You are doing what? No, don't do that -- the law considers
it a misrepresentation, even if it does no harm to others." Having
received such guidance, I am responsible for implementing it even
if the board never sets a policy for the ASF as a whole.
So, why didn't I apply the changes to httpd already? The answer is
because I am waiting for a board decision, and because the httpd source
contains so much collective work that it is very hard to find a
file that cannot be argued as (at least) a joint effort by the ASF.
That is not the case for Apache Jackrabbit. Much of the Jackrabbit
code was developed long before it was licensed to the ASF in a
CCLA+grant. Furthermore, all of the contributions since then have
been under CLAs. Therefore, I do know the copyright owner of those
files and I do know what can't be said in the notices. Even so,
I put off the change until the prep for our formal FCS 1.0 release
made it necessary.
On Apr 21, 2006, at 11:46 AM, William A. Rowe, Jr. wrote:
This just isn't making sense as I read the svn tree for jackrabbit
however.
Individual files contain -no- copyright, -no- indicication of where
the
copyright claim is
http://svn.apache.org/repos/asf/jackrabbit/trunk/jackrabbit/src/
main/java/org/apache/jackrabbit/
has JcrConstants.java pointing to http://www.apache.org/licenses/
LICENSE-2.0
and claiming the license but no copyright. That LICENSE.txt is
bundled in the
tree (not, you'll note, LICENSE-2.0) at that level. NOTICE.txt
tells us...
This product includes software developed by
The Apache Software Foundation (http://www.apache.org/).
Based on source code originally developed by
Day Software (http://www.day.com/).
which is all well and good, but doesn't assert copyrights.
The only way to "assert copyright" is to accuse someone of infringing
those exclusive rights. A notice is not an assertion -- it is supposed
to be a simple statement of fact.
It's not until you climb all the way up to
http://svn.apache.org/repos/asf/jackrabbit/trunk/jackrabbit/
(outside of even the src/ tree!) that you discover...
http://svn.apache.org/repos/asf/jackrabbit/trunk/jackrabbit/README.txt
The source tree is trunk. The "src" tree is a directory that Maven
uses to look for java source code. The README is included in all
of our release jars. Thus, the ASF collective work copyright notice
is present in all of our work products and our website.
I'm really completely unclear how this protects the files we
author, the files
authored by others (which we have appropriately appropriated) and
the files on
which no copyright is claimed (e.g. apr/ examples public domain.)
That is irrelevant. They are protected by copyright law, regardless
of the notice or lack thereof.
Jackrabbit is the test animal (so to speak). Let's give the board
time to consider what, if anything, should be done for general policy.
I don't know of any reason to hurry a 2.0.x release (heck, I don't
know of any reason to continue its development), but I also don't think
releasing one with modified copyright years is any more or less legal
than continuing to distribute the previous releases. Nor do I think
any of our users will ever care enough to challenge the notices.
Personally, I think it is part of the board's job (and the chairs,
being their conduit) to take care of this legal mumbo jumbo and thus
save the developers from having to worry about it. It simply takes
a lot longer than expected, sometimes, because nobody likes to do
this kind of work. We should not expect it to be any less bug-free
than other procedures.
....Roy
