[Savannah-register-public] Re: [gnu.org #343410] Licensing GPL/GFDL: Perl code and embedded documentation

[Sylvain Beucler]

On Sat, Sep 15, 2007 at 11:52:13AM -0400, (Jari Aalto) via RT wrote:
> The application's distribution archive includes structure:
> 
>     doc/txt/program.txt
>     doc/html/program.html
>     doc/man/program.1
> 
> to where the Makefile can extract the POD section's content using
> various Perl language filters:  man, text, html respectively.
> 
> The Savannah admin has asked to license these generated output files
> under GFDL.

Hi,

Actually we need the _source form_ of the documentation - that is, the
POD - to be licensed under the GFDL (or optionaly dual-licensed
GPL/GFDL). Generated documentation will be under the same
(dual)license, as it's a derived work of the source form.

More generally:

> 1. What is the effect of the extracting process that uses Perl's own
>    commands to generate the content in license wise?  Does it affect
>    the final output's licensing conditions?

In this case the generated output is a derived work of the input -
input's license applies.

> 2. Can the extracted content be put under Dual GPL/GFDL licence or
>    does it share the cracteristics of the source where it came from?

The input's license still applies, so the output's license needs to be
compatible with the input's.


I think there are a few particular cases where you can consider the
output as an entire new work. This was used to generate the first free
windows.h (ancestry of mingw) using a free source code that included
<windows.h> and displayed the values of some specific variables; this
free source code could be compiled to produce a program that could
produce a .h with the same values. But as you can see this is really a
corner case and this uses a small subset of the input under conditions
that are compatible with the input's. In your case the generated
documentation is clearly a derived work of the source form.

-- 
Sylvain