[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Please provide source code for Tandem Repeats Finder under a freely redistributable license



Hi Yozen,

On Thu, May 12, 2016 at 08:11:53PM -0400, Yozen Hernandez wrote:
> Dear Andreas,
> 
> Thank you for contacting about packaging TRF. I indeed missed your email.
> 
> Unfortunately, the source code for TRF is not freely available. I do not
> know the Debian Free Software Guidelines well enough to decide whether or
> not this is allowed,

The Debian Free Software Guidelines[1] are widely accepted as Open
Source definition and it is actually that the source code is freely
available.

> but I may have a solution that I have seen other
> packages use. For instance, the Oracle Java packages I have seen ask the
> user to accept a license agreement. The package script then goes on to
> download the appropriate binary upon acceptance. This could be part of the
> RepeatMasker package since TRF is required by that software, and I can
> assist by providing the URLs for the i386 and x64 builds.

There are several glitches in this suggestion:

   1. The Oracle Java packages are not part of Debian.  Debian ships
      with OpenJDK.  In so far the comparison is weak.
   2. For official Debian packages there is no point to ask for a
      license since the user can be sure that the installed software
      is per definition free.
   3. Debian packages can only depend from other Debian packages
      but not from unofficial ones like Oracle Java or your suggested
      TRF
   4. Debian ships for several architectures not only i386 and x64 (in
      Debian nomenclature amd64).  For instance architectures like arm64
      and ppc64 might become quite interesting in the near future and
      we try to care for building on these architectures as well

So far for the general considerations.  In the specific case of
scientific software we also consider the free availability of the source
code as very important to prove the correctness of the results and
enable reproducibility.  So in this case there are extra good reasons
for publishing the code.
 
> Another work around could be to ask the user to manually download and
> install TRF themselves by providing them with the URL to our downloads page.

As said above a Debian package is not allowed to depend from external
resources.
 
> I understand that these are not ideal solutions, but for the forseeable
> future we will not license TRF under a free software license.

I have no idea whether this might be open for discussion at your site
and what might be the motivation to keep the code of a scientific tool
closed.  In case you might like to re-think it I'd like to point out
the following advantages:

Due to the work of the Debian Med team Debian and its derivatives like
Ubuntu gained quite some coverage in biological research.  Providing TRF
via Debian packages does not only simplify the installation and
maintenance for users (you might keep cluster installations in mind).
 
The fact that we are providing so called metapackages depending from all
biological applications installing any application you see on this so
called tasks page[2] might bring TRF automatically on users computers
who might become aware of your tool just because it is included here and
you might gain additional users of your software.

As you can see on the tasks page[2] we also put some importance on
specifying the according publications which gives the authors of the
software some extra credit.

I could give several more good reasons but I don't know your motivation
to keep the source code closed and may be you could rethink the decision
by including the arguments above.

Kind regards

       Andreas.

[1] https://www.debian.org/social_contract#guidelines
[2] http://blends.debian.org/med/tasks/bio

-- 
http://fam-tille.de


Reply to: