Simplify openjpeg copyright

[gcode-importer]

Originally reported on Google Code with ID 318

Right now openjpeg copyright is a mess:

http://metadata.ftp-master.debian.org/changelogs/main/o/openjpeg/unstable_copyright

251 lines for a BSD-2 license with tiny copyright differences.

Reported by malaterre on 2014-03-27 07:51:55

[gcode-importer]

Well, the thing is that some people contribute to JPIP, other to JPWL, other to MJ2,
etc. 

The differences are in the copyright, not the license (which should be always 2-clauses-BSD).

Ideas/suggestions on how to manage growing copyright notices in open-source projects
welcome.

Reported by detonin on 2014-03-27 09:46:22

[gcode-importer]

Eg:

https://wiki.python.org/moin/PythonSoftwareFoundationLicenseFaq#What_if_I_want_to_contribute_my_code_to_the_PSF.3F

Contributor should donate copyright to some kind of foundation (OpenJPEG Foundation?).

Or else switch to Apache 2.0, which simplify the BSD handling.

Reported by malaterre on 2014-03-27 09:52:38

[gcode-importer]

Here is my proposition. Everyone who contribute to OpenJPEG should donate their copyright
to:

  Copyright (c) 2002-2014, OpenJPEG contributors

Another alternative could be (if we want to keep the first contributor):

  Copyright (c) 2002-2014, Universite catholique de Louvain (UCL), Belgium
  Copyright (c) 2002-2014, OpenJPEG contributors

If we are ok with this, we can start pinging people to reply to this thread and have
their consents.

Reported by malaterre on 2014-04-08 10:52:24

[gcode-importer]

+1 for the solution:
  Copyright (c) 2002-2014, Universite catholique de Louvain (UCL), Belgium
  Copyright (c) 2002-2014, OpenJPEG contributors

Here is the list of the current contributors that have their name in one or several
copyrights : 

Benoit Macq
David Janssens
Kaori Hagihara
Jerome Fimes
Giuseppe Baruffa
Mickaël Savinaud 
Mathieu Malaterre
Yannick Verschueren
Herve Drolon
Francois-Olivier Devaux
Antonin Descampe
Centre National d'Etudes Spatiales (CNES), France
CS Systemes d'Information, France
Parvatha Elangovan
Jonathan Ballard
Callum Lerwick

Did I forget someone ?
Who shall we contact for CS and CNES ?

Reported by detonin on 2014-04-10 12:03:24

[gcode-importer]

For reference, this happen to other project: https://github.com/openseadragon/openseadragon/issues/58

Reported by malaterre on 2014-04-10 12:17:32

[gcode-importer]

I understand the simplification, you have my permission.

Reported by malaterre on 2014-04-10 12:18:07

[gcode-importer]

You have my permission as well.

Reported by detonin on 2014-04-10 12:24:48

[gcode-importer]

Reported by malaterre on 2014-04-10 12:26:29

[gcode-importer]

Reported by malaterre on 2014-04-10 12:26:42

• Labels added: Priority-High
• Labels removed: Priority-Medium

[gcode-importer]

you have my permission as well.

Benoit Macq

Reported by Benoit.Macq on 2014-04-10 13:49:15

[gcode-importer]

I do not agree with the simplification as stated above with the BSD license that has
two copyright lines, one for the university and one for contributors. I believe an
improved choice is given in the form of the Apache 2.0 license since it is more effective
in desired simplification with compatibility to other licenses, licensors, and assignees.
The BSD license with more than one assignee is effectively the MIT license, which puts
the terms within educational bounds. In either case, the BSD license and MIT license
limits usage outside of educational code. That is often overlooked. We should allow
open usage of the source code instead of further impede application.

Jonathan Ballard

Reported by dzonatas on 2014-04-10 18:15:17

[gcode-importer]

Hello everybody !
It's nice to see that the OpenJPEG projects is developing so well. Congratulations
to all.

I understand the simplification proposed here. You have my permission !

Reported by fodevaux on 2014-04-11 08:14:46

[gcode-importer]

@dzonatas: AFAIU, the reason why you don't agree with the proposed simplification is
that in your opinion the Apache 2.0 license is better than the BSD and OpenJPEG should
adopt it. But this is another debate. The primary purpose of the proposed simplification
is only to ease the work of, for instance, debian package managers when they have to
gather all the different copyrights from all the files in the project. It would be
so much easier to have the same license everywhere and AFAIK it does not change any
right anyone might have on the source code, compared to the current situation. If we
agree that a change from BSD to Apache is another (relevant) issue, do you think you
could reconsider your opinion on this very topic ? Let us know what you think. 

Thanks, 

Antonin.

Reported by detonin on 2014-04-11 13:44:11

[gcode-importer]

The word "better" is not my opinion in regards to acceptance of Apache 2.0 license.
The purpose of the Apache 2.0 license suits the needs in this topic situation. I compromised
my own personal choice and refrained from further suggestion of my thought on superior
licensing terms. I have no doubt that Debian maintainers will also accept Apache 2.0
licensing terms, as it is compatible with Debian Committee's choice of licenses.

I'm in favor of the single assignee if each contributor agrees that our name(s) will
be kept in (to be attached) contributor file along with opt-in choice of accreditation.
I am certain that meets this topic demand. For sample, in mean time development of
DWT optimizations, I accredit: American River College.

Jonathan Ballard

Reported by dzonatas on 2014-04-11 14:34:18

[gcode-importer]

@dzonatas Ok so in summary, you are ok with BSD single assignement:

  Copyright (c) 2002-2014, OpenJPEG contributors

Even if this was not explicitly stated, we are of course keeping the AUTHORS file as-is.
If we missed anyone, here is your time to state it loud.

@antonin, do you know how hard it will be to have someone from Universite catholique
de Louvain (UCL), Belgium to agree to the copyright change ?

Reported by malaterre on 2014-04-11 14:40:15

[gcode-importer]

Hi to all,

I understand the need for license simplification. 
You have my permission. 

Hervé Drolon
FreeImage Team

Reported by FreeImage on 2014-04-11 20:54:17

[gcode-importer]

Hi,
I can't give my permission for this copyright modification. For different reason:
Some of my work have been done under a contract with CNES so you need their permission
for this part of my work. In this case I am just author not the copyright owner. It
is the same for some work done during my CS-SI contract, the copyright owner is my
company. For the work done during my personal time, I am the copyright owner. So it
is quite difficult to give my permission.
Moreover the first point of the bsd claim that the above copyright notice could not
be modified. 
The best way I think is to move to an Apache license and/or create steering committee
attach to a foundation with agreement acceptation before committing. You can find a
good example in the project Orekit (https://www.orekit.org/forge/attachments/338/OREKIT_Governance.pdf)

 Best

Reported by savmickael on 2014-04-13 22:02:04

[gcode-importer]

@savmickael there is no difference at all whether we move to BSD asking you to change
the copyright change (this is legal for CS representative) or move to Apache License.

In both case someone with CS representation need to give consent.

Reported by malaterre on 2014-04-14 07:07:13

[gcode-importer]

@dzonatas: Thanks for the effort you made to compromise your personal choice.
@mathieu: I do not know how hard it would be to have UCL agree on this. I check.
@savmickael: Moving to Apache license does not solve the copyright simplification issue
we're trying to fix here. Or am I missing sth ? 

Reported by detonin on 2014-04-14 08:49:56

[gcode-importer]

Answer from UCL regarding the single assignee proposition is that as initiator, host,
maintainer, funder, and main contributor of this project they want to keep their name
in the copyright. By the way, this is what is done in many open-source projects, like
OREKIT, mentioned by savmickael, which uses a NOTICE file to credit contributions (https://www.orekit.org/forge/projects/orekit/repository/revisions/master/entry/NOTICE.txt).

@dzonatas: do you think you could live with the 2-assignees proposition and if not,
explain a bit more why do you think it would be bad/unfair ? Thanks.

@savmickael: do you know who we are supposed to contact to ask permission regarding
CS and CNES copyright ? 

Reported by detonin on 2014-04-15 15:40:02

[gcode-importer]

If UCL agreed to the single assignee, then we still need the single assignee from UCL.
"Universite catholique de Louvain (UCL), Belgium" does not constitute the single assignee
required by this BSD license.

Reported by dzonatas on 2014-04-16 02:30:52

[gcode-importer]

@dzonatas: I'm afraid I did not understand what you meant by "single assignee". Could
you explain why "Universite catholique de Louvain (UCL), Belgium" does not constitute
a valid assignee for BSD ? And if you would agree with the proposition in comment #4,
knowing the UCL position on that topic ? And if not, could you write down what a suitable
solution would be ? Thanks again. 

Reported by detonin on 2014-04-16 09:46:02

[gcode-importer]

@antonin: In comment #20, as stated, "they want to keep their name in the copyright."
The name given by UCL is only their accreditation, as demonstrated in #14. If nobody
else on the copyright is part of UCL, then that is the defect.

Reported by dzonatas on 2014-04-16 14:26:27

[gcode-importer]

@dzonatas: the copyright can be owned by the university on not necessarily by a physical
person. Then, in the AUTHORS, or THANKS, or NOTICE or whatever file, we put the names
of the people having effectively written down the code. It does not appear to be a
defect to me. Please explain further.

Reported by detonin on 2014-04-17 07:53:45

[gcode-importer]

Copyrights are of the mind. We did not agree to the "UCL license."

Reported by dzonatas on 2014-04-17 13:51:14

[gcode-importer]

If we include an ARTIST file, we could represent the UCL copyright upon an additional
entity denotation, from unicode: http://unicode-table.com/en/sections/enclosed-alphanumerics/
( @antonin )

Sample solution: The latin-"capitol"-circle A would then be an entity as single assignee
for this artistic representation TO BE ASSIGNED (or underwritten) as: "Copyright (c)
2002-2014, Universite catholique de Louvain (UCL), Belgium"

An ARTIST may insert the denotation for claims apart from BSD-AUTHORS.

This avoids an explicit reassignment against desired intention stated in #20. The alternative,
obvious to (BSD)UNIX folk, is fork() the front-end code.

[Down-Note: I do not consider QR-CODED condition statements in C-code as art.]

Jonathan Ballard

Reported by dzonatas on 2014-04-17 19:42:41

[gcode-importer]

In further research:

(1) In regards to OreKit, we have STEM, which has been mirrored globally:
www.whitehouse.gov/sites/default/files/microsites/ostp/stem_stratplan_2013.pdf

(2) Understanding STEM, we MAY enable default assignment by entities Ⓢ, Ⓣ,
Ⓔ, and Ⓜ  for any claim or likewise rather than rely on single assignment.
I consider this the higher-level fork() without confusion of the question
"which university" is of mind.

(3) I neither opened nor looked at the kakadu source code despite of me
being held responsible for this code, and in spite of me unjustly and
relentlessly being called patent troll. I still have the original mail
message from one contracted OpenSim/SL contributor that linked me the
source code. I kept the linked zip file as evidence. In reality, the
lifting-scheme is minor in comparison to the patented cellular approach.

(4) The patented cellular approach is inevitable (to be implemented) with
appropriate levels of parallelism. Imagine how ⓈⓉⒺⓂ can help prevent
further division as in this issue.

(5) If we prevent further division, we SHOULD consider this issue an
enhancement rather than the defect as classified.

As the scientist, I neither judge, covet, nor indoctrinate us nor our
course of action. It is not me that any particular person does not
understand, yet (outside campus) I can refer back to (1).

[I earned my cup of coffee today.]

Reported by dzonatas on 2014-04-18 12:31:23

[gcode-importer]

@dzonatas: can we try to keep the solution (and the debate) as simple as possible ?
Could you state if you eventually agree with the solution hereunder ?

  Copyright (c) 2002-2014, Universite catholique de Louvain (UCL), Belgium
  Copyright (c) 2002-2014, OpenJPEG contributors

Many thanks.

Reported by detonin on 2014-04-24 12:37:47

[gcode-importer]

Hello, for us (DSPLab - University of Perugia) it is OK to move to the new copyright
denomination suggested by Antonin.
Giuseppe

Reported by gbaruffa on 2014-05-01 08:31:47