cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ralph Goers <>
Subject [Fwd: Re: [VOTE] Removing author tags]
Date Tue, 03 May 2005 01:19:27 GMT
Sorry if this shows up twice, but I posted it two hours ago and it still 
hasn't hit the list...

-------- Original Message --------
Subject: 	Re: [VOTE] Removing author tags
Date: 	Mon, 02 May 2005 16:28:11 -0700
From: 	Ralph Goers <>
References: 	<> 

Daniel Fagerstrom wrote:

> Sylvain Wallez wrote:
>> So I propose to remove @author tags with people names from all our 
>> source files.
> +1
Well, I'm not really sure how to vote on this.  I've pasted an email 
from Larry Rosen to the Apache Legal list that seems to say that the 
Cocoon source files should not have the Apache copyright in them, but 
instead have the copyright of the committers who made the contibution.  
Obviously, this is in direct contrast to what we are doing and frankly, 
I think its nuts. 

While this doesn't have a lot to do with the @author tags, per se, it 
does seem to indicate that the files should have a "record of ownership".

So before I add my +1 I'd at least like to know if our policy of 
attributing the copyrights to Apache and not the authors needs to be 
changed (at which point the @author tags could go away anyway).


List-Help: <>
List-Unsubscribe: <>
List-Post: <>
List-Id: <>
Delivered-To: mailing list
From: "Lawrence Rosen" <>
To: <>, <>
Subject: RE: Copyright text, and javadoc license
Date: Sat, 5 Mar 2005 08:24:06 -0800

> Has there been any progress on the issue of
>   "Copyright [yyyy] The Apache Software Foundation or its
>   licensors, as applicable"
> vs
>   "Copyright (c) 2001-2004 - Apache Software Foundation"?

During the past few weeks Robyn Wagner, Jennifer Machovec, Jim Barnett and I
discussed this issue via email. I drafted for their review an email that I
proposed to send to legal-discuss to summarize the legal requirements for
copyright notices on Apache works; I now step up to the plate and copy that
summary below. Both Jennifer and Jim agree that the summary is legally
correct, but they worry that the Apache members "want" to do it differently.
I have suggested that lawyers should tell clients what the law is, but it is
up to the clients to decide whether to follow the law. So here it is (with
apologies for its length) for your review and discussion:


There have been concerns raised recently about copyright notices on software
distributed by ASF. Here's basically what the law allows: We can place an
ASF copyright notice on any original works of authorship authored or owned
by the Apache Software Foundation, and nowhere else. Since we do not author
contributions as such - our Contributors do! - which original works of
authorship can actually bear the ASF copyright notice?

An ASF copyright notice can be placed only (1) on the ASF website itself and
on related expressive pronouncements of the ASF board of directors, its
officers or agents created on behalf of ASF and published to the world as
ASF's "voice"; (2) on the CVS data bases containing the Contributions
collected, selected and arranged in accordance with ASF-authorized
processes. The former are generally *original work* copyrights; the latter
are *collective work* copyrights. (A more general term for collective works
under US copyright law would be a *compilation work* but that's not an
important distinction for this thread and would be confusing to all
programmers.) In some cases, third parties or the ASF itself may create
modifications of (1) and (2). Only if ASF itself does the modification may
it (3) place its *derivative work* copyright notice on that work. See 17 USC

In all such cases, the copyright notice for ASF's copyrightable works
published this year would be:

Copyright 2005 Apache Software Foundation.

The form of the ASF copyright notice is specified in 17 USC 401 to include
just three components: the word "Copyright", the year of first publication,
and the name of the owner of the copyright in the work.

It is never necessary to identify in the notice itself whether the claimed
copyright is for an original work, a collective work, or a derivative work.
Those details are not for the notice; instead, they are specified in an
application for registration ASF may file with the Library of Congress - and
which we must file if ASF ever intends to enforce its copyrights in federal
court. (The topic of registering ASF's copyrights is reserved for another

The confusion on this list has been over the collective (or compilation)
work copyrights that Apache can rightfully claim in its collection of
Contributions. Merely by collecting Contributions according to established
and intelligent procedures, ASF has made a sufficient level of creative
input to claim a collective work copyright in the entire collection of ASF

The law is clear, however: ASF's copyright in a collective work does not
extend to the preexisting materials placed there by its Contributors. 17 USC
103. Whoever owns the copyrights in those Contributions retains that
ownership despite their licensing it to ASF for inclusion into ASF's
collective works. Some attorneys refer to the copyright in a collective work
as "thin" because its owner can only prevent infringement of the collective
work itself, not of the individual components. Even thin copyrights can be
very valuable, however, because of the additional effects of trademark
protection (e.g., the Apache brand) and the persuasive authority of a
respected board of directors and the community of Contributors it

Every Contributor should be encouraged to place appropriate copyright
notices in the source code of its own copyrightable works at the time of its
contribution. ASF is not responsible for doing so, nor is it responsible for
the accuracy of those notices placed in source code by its Contributors. The
Apache Software License effectively disclaims the warranty of
non-infringement, so other people's copyright disputes are not our legal
concern. Furthermore, ASF's Contributor Agreements protect ASF from

Any downstream recipient of ASF's collective works takes those works subject
to all upstream licenses even if the only copyright notice appearing on the
work is that reflecting the ASF *collective work* copyright. Any downstream
recipient is on notice that the software cannot be copied except under the
terms and conditions specified in the Apache Software License. The ASF
website contains all important license notices and text. The source code
with the original copyright notices supplied by the Contributors is
available on ASF's website for all to see. No responsible commercial entity
who mattered would dare claim ignorance of Apache's license and processes.

ASF's copyright notice is all the notice needed, because our copyrightable
work is available only under the terms of our license, which explains the
rules for all.

ASF will not remove (indeed, should not remove) any pre-existing copyright
notices. This rule applies to contributed source code only. Downstream
licensees are entitled to that information so they can know their rights and
obligations, particularly if they use the source code to create derivative
works. But there is no requirement to publish the original copyright notices
when "compiling" source code into executable code or when distributing
executable versions of software. Such transformations are derivative works
owned by ASF or by the person who performed the transformations, and deserve
a new copyright notice.

Once again, the only legitimate places where ASF may place its copyright
notice are on its website, on its publications, and on its collected data
bases of contributed software. In the latter situation, ASF may claim a
collective work copyright by virtue simply of having collected it. In all
such cases, the only legitimate ASF copyright notice is:

Copyright 2005 Apache Software Foundation.


I also copy below the email exchange between Jennifer and me about the
expressed desire of Apache members to remove copyright noticed placed by
Contributors into their contributions for Apache. That correspondence is
copied below, with Jennifer's note at the bottom and my response above it:

Hi Jennifer,

Thanks for your input and for explaining the history that guided your
previous decisions.

You are absolutely right that it would be improper to place an ASF copyright
notice on files that are "contributed in their entirety and are not modified
by ASF or other contributors." What I don't then understand is why the ASF
should want to change those existing copyright notices at all. Just leave
them alone.

Regardless of the "*very* strong feeling" by ASF members that they want to
highlight the ASF identity rather than that of the contributors, modifying
or removing existing copyright notices in source code of contributions is
not the way to do that. It has the great disadvantage of obscuring the
provenance of the code, and it can lead to misunderstandings and grievances
later if contributors should want to use their own code in other ways. No
matter how adamant the members and board are, it is appropriate for lawyers
like us to say "you shouldn't be doing that."

If copyright owners like BEA and IBM acquiesce in the removal of their own
copyright notices, they may be denying themselves the full protection of the
copyright law. (See 17 USC 405(e).) And they needn't do so as long as a
*valid* Apache copyright notice is placed on the entire collective work.
(See 17 USC 404(a).)

You were concerned about the location of the Apache copyright notices on the
collective work. I'm not sure why that's become such an issue. The Apache
website is where that code is made available. What's wrong with a standard
Apache copyright notice on every such download page?

I'm certainly willing to take this issue back to ASF for reconsideration. If
I didn't feel strongly about this point I wouldn't have spent as much time
as I already did trying to explain it. My objective in reviewing the draft
with you all first was to see if perhaps we attorneys could present a common
front at least as to the interpretation of the law. Then, as we all know, if
the client still insists on doing foolish or bizarre things, it isn't our


View raw message