Skip to content

Copyright versus Licenses

3 messages · Bryan McLellan, Simon Urbanek, Thomas Lumley

#
My company recently started using a R library from RCRAN that is
licensed under the LGPL Version 2 or greater per the DESCRIPTION file,
but contains no copy of the LGPL notice, or any copyright notice. I've
grown accustomed to paying attention to copyright and licensing as a
Debian package maintainer, and sent the author of the package an email
expressing my concern. The author believed that assigning themselves
copyright was incompatible with licensing the library under the terms
of the LGPL. I disagree, and further contend that without copyright
notice, the [copyright] license loses a certain degree of
applicability, as it becomes inconclusive as to who is licensing the
software under the terms of the LGPL. Not knowing who I was, the
library author asked me to start a discussion of the subject on this
list, presumably so they could see the opinions of others that they
trust.

The LGPL itself [1], in the final section entitled "How to Apply These
Terms to Your New Libraries", the license provides a template for
adding to the top of each source code file that contains a copyright
line, a general notice regarding the lack of warranty, and information
on where to obtain a full copy of the license. The GPL HOWTO [2]
expresses similar instructions for the inclusion of a copyright line
with the license. I know that R distributes copies of common licenses
under 'share/licenses' in the R source. Debian does as well in
'/usr/share/common-licenses/' for the purpose of not having to include
the full LICENSE and/or COPYING file with every package that uses a
common open source license, allowing the use of verbage such as "The
Debian packaging is ? 2010 [author] and is licensed under the Apache
License version 2.0. On debian and derived systems, see
`/usr/share/common-licenses/Apache-2.0' for the complete text." The R
manual for writing extensions suggests a similar approach to avoiding
duplication in Section 1.1 [3].

The R manual for writing extensions also mentions [4] in Section 1.1.1
the optional use of a Copyright field in the DESCRIPTION file,
separate from the License field. As this section implies that the
DESCRIPTION file format is based on the debian control file format, I
assume the goal is to keep these lines simple, generally under 80
characters do to average terminal width. As such, I don't assume this
field is recommended for complete copyright information for a library
with multiple contributors. The aforementioned article does not
specify where a developer should alternately put copyright
information, perhaps assuming one would add it to each source code
file as is recommended by GNU.

In closing, do the R developers believe that including a Copyright
notice is imperative with a Copyright License?

If so, what advice do they have for those writing and contributing
open source R libraries as to where this notice should go?

Should that information perhaps be added to the R manual for extensions?

Bryan McLellan

[1] http://www.gnu.org/licenses/lgpl-2.1.txt
[2] http://www.gnu.org/licenses/gpl-howto.html
[3] http://cran.r-project.org/doc/manuals/R-exts.html#Package-structure
[4] http://cran.r-project.org/doc/manuals/R-exts.html#The-DESCRIPTION-file
#
On Jan 18, 2010, at 23:06 , Bryan McLellan wrote:

            
I suspect you meant R package as R libraries have no DESCRIPTION ...
[Disclaimer: I'm not a lawyer and this is not a legal advice. I'm not  
representing anyone but myself.]

Copyright is the right that the author of an original work holds  
automatically (unless someone else can claim to own his work - e.g.  
his employer etc.) under the Berne Convention. The copyright gives  
only the author all rights - including but not limited to the right to  
copy, modify, distribute the work etc.  Licenses are used to give some  
of those rights to other people under certain conditions. Hence you  
cannot "assign yourself copyright" because you already have it (and if  
you don't then your cannot assign it). Also you don't need to give the  
"copyright" to anyone else - you can give certain rights to others  
using licenses -- such as GPL, LGPL, EUPL etc. -- but you don't give  
copyright by those since you have far more rights as the author (e.g.,  
you can do whatever your want with the original work beyond the  
restrictions of the license). There are cases where you may want to  
give your copyright to someone, e.g., an organization that represents  
all authors of a project which makes it easier to change licenses  
etc., but that is a different story.
I think that is a good practice since it makes it easy to find the  
copyright holder in case of licensing questions.

However, I suspect that the Author field in the DESCRIPTION could be  
interpreted as the copyright holder unless anything else is specified  
(if the author doesn't have copyright s/he would have to specify the  
Copyright: field). Sometimes single-author packages just state the  
author and license assuming that it is clear enough. For large  
projects with many copyright holders you really want to mark the  
individual files. In either case, marking individual files is clearer  
as they may be available separately. I don't think there is a strict  
rule - if tested by the law I suspect that the intent is the most  
important factor.

This is just my $0.02 -- it would be a really great if authors  
understood exactly what is a copyright and what is a license.  
Unfortunately there is a lot of confusion as people read about certain  
details (e.g. from GNU FAQ) without understanding the underlying  
mechanisms and the implications.

Cheers,
Simon
#
On Mon, 18 Jan 2010, Bryan McLellan wrote:

            
Like Simon, I'm speaking only for myself here.

Not having an explicit copyright statement certainly makes it harder for people downstream who have questions, so explicit statements seem like a good thing.
Yes, in the source code is good.

In addition, if you look at the source code for R, the approach there has been to have a file COPYRIGHTS that lists the original copyright holders of bits of the code.  This seems like a good idea if there are multiple sources.  It's especially useful if there are areas of doubt and uncertainty (eg code published on ftp sites before people worried about licenses), since it at least gives you rigidly defined areas of doubt and uncertainty[5]
-thomas

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle



[5] Adams D, (1978) Hitchhiker's Guide. BBC Radio.