db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jennifer B Machovec <macho...@us.ibm.com>
Subject Re: Derby code copyright question
Date Mon, 27 Sep 2004 14:27:11 GMT
There is a significant difference between having a contributor give ASF 
and its end-users a broad copyright and patent license grant (which does 
not require ASF to distribute the code under any particular licensing 
terms) and telling the contributor that he must assign away all rights of 
ownership.  Under the process that has been in place at ASF since its 
inception, a contributor can continue to use or license his code in any 
manner that he sees fit (concurrently with ASF's use of the code under the 
license grant) because he retains ownership of the intellectual property. 
Asking the contributor to transfer ownership to ASF would destroy these 
rights.  I imagine that a significant number of contributors would be 
reluctant to give up their rights in their creation entirely.

As far as the issue of additional copyright holders, there can very well 
be several copyright owners affiliated with a particular project.  ASF, 
for instance, generally holds copyright for the collection of works that 
comprise the project.  Developers who modify ASF code hold copyright in 
the derivative works thus created.  Even though third-party copyright 
notices may not be included for all contributors, developers and companies 
who have submitted code to ASF for the last several years have 
nevertheless retained the status of copyright owner for their code.  This 
doesn't thwart ASF from being able to demonstrate a clear provenance in 
the IP; as long as those contributors have executed CLAs or otherwise 
submitted code under documents containing broad license grants similar to 
those in effect today, the scope of those grants should be sufficient for 
ASF to keep doing business as usual.



"Geir Magnusson Jr." <geirm@apache.org>
09/27/2004 09:59 AM
Please respond to "Derby Development"
        To:     "Derby Development" <derby-dev@db.apache.org>
        Subject:        Re: Derby code copyright question

On Sep 26, 2004, at 6:20 PM, Daniel John Debrunner wrote:

> Hash: SHA1
> Geir Magnusson Jr. wrote:
>> On Sep 26, 2004, at 11:37 AM, Daniel John Debrunner wrote:
>>> Hash: SHA1
>>> Geir Magnusson Jr. wrote:
>>>> Traditionally, the only time we distribute code under someone else's
>>>> copyright is when we are including other works as a convenience for 
>>>>  our
>>>> users, such as packaging something in a distribution.  Otherwise, we
>>>> tend not to host projects owned by other entities.  Sourceforge is 
>>>> good
>>>> for that.
>>> Incorrect.
>> Not "Incorrect".  Show me software in the ASF that is not (c) ASF,
>> besides standard APIs or such.  We may have attributed (c) ASF by
>> mistake, or incorrectly, but it is all (c) ASF except for inclusions 
>> of
>> outside works for packaging convenience.
> Geir, I think we must be misunderstanding each other.
> All code distributed by ASF may only have (C) ASF notices but that does
> not change the fact that the code does have additional copyright
> holders. I thought you were claiming that ASF only distributed code
> where the only copyright holder was ASF.

You can't really have additional copyright holders, if I understand 
copyright law correctly.  You can have licensees.  I'll be the first to 
admit that I had this wrong before this came up.  I thought that the 
ASF had free and clear copyright ownership, and the contributor did 
too.  Thus, the ASF was free to manage the code in the way that we saw 
fit - for example, we could change the license under which that code 
was distributed - the license for anything and everything in the 
distro.  I don't believe we can do that with derby if it's (c) IBM 
unless we get another agreement from IBM allowing us to do that.  I 
don't want to go down that road because of the administrative burden 
that would place on the ASF.

I'm just baffled why this is such a problem - why can't IBM just assign 

> Derby wants to follow existing ASF practices, I'm now no longer have 
> any
> idea what folks are objecting to. NOTICES file seems to be the new
> issue, but again Derby wants to follow existing ASF practice, as seen 
> in
> the Xerces Java notice file and the Beehive notice file.

I guess the question I have, looking at Xerces, and I as a non-lawyer 
read it this way

"Xerces is (c) ASF in its entirety, and we're giving credit to IBM and 
Sun for contributions to this work that were *originally* (c) IBM or 
(c) SUNW"

The implication to this non-lawyer was *originally* (c) IBM but isn't 
now - it's wholly (c) ASF.  So if I have a problem, I talk to ASF.  I 
don't need to rope in Sun and IBM.

I think that it serves everyones interest to keep things clear and 
simple.  One of the things we strive to do at the ASF is provide 
software with clear IP provenance that doesn't require our users to 
deal with any other organization for any issues surrounding the 
distributed work.

I think everyone involved in this issue feels that is a worthy 
objective of the ASF, so I just don't grok the pushback we're getting 
on this.


Geir Magnusson Jr                                  +1-203-665-6437

View raw message