Can autotalent be packaged for Fedora?

[David Cornette]

I am thinking of packaging autotalent
<http://web.mit.edu/tbaran/www/autotalent.html> for Fedora, but I first wanted
to make sure that license and copyright permitted it.

The source distribution contains the following files:
autotalent.c
COPYING
ladspa.h
Makefile
mayer_fft.c
mayer_fft.h
README

The README file says "By Thomas A. Baran", but does not say anything about
copyright or license.  The file COPYING states the following, which is followed
by the text of the GPLv2:

The bulk of the code in Autotalent is released under GPL2.  However,
the FFT routine was taken from PureData
(http://crca.ucsd.edu/~msp/software.html), which was released under a
license that is similar to the BSD license.  So with the exception of
the mayer_fft.* files, everything should fall under GPL2.

-Tom

END QUOTE

The files autotalent.c and Makefile both state that they are under GPLv2+.
The file ladspa.h is under LGPLv2.1+

The file mayer_fft.h has no attribution, copyright or license information at
all.  The file mayer.c has this to say:

/* This is the FFT routine taken from PureData, a great piece of
software by Miller S. Puckette.
http://crca.ucsd.edu/~msp/software.html */

/*
** FFT and FHT routines
**  Copyright 1988, 1993; Ron Mayer
**
[Documentation of functions and their signatures omitted]
**
**
** NOTE: This routine uses at least 2 patented algorithms, and may be
**       under the restrictions of a bunch of different organizations.
**       Although I wrote it completely myself, it is kind of a derivative
**       of a routine I once authored and released under the GPL, so it
**       may fall under the free software foundation's restrictions;
**       it was worked on as a Stanford Univ project, so they claim
**       some rights to it; it was further optimized at work here, so
**       I think this company claims parts of it.  The patents are
**       held by R. Bracewell (the FHT algorithm) and O. Buneman (the
**       trig generator), both at Stanford Univ.
**       If it were up to me, I'd say go do whatever you want with it;
**       but it would be polite to give credit to the following people
**       if you use this anywhere:
**           Euler     - probable inventor of the fourier transform.
**           Gauss     - probable inventor of the FFT.
**           Hartley   - probable inventor of the hartley transform.
**           Buneman   - for a really cool trig generator
**           Mayer(me) - for authoring this particular version and
**                       including all the optimizations in one package.
**       Thanks,
**       Ron Mayer; mayer(a)acuson.com
**
*/

/* This is a slightly modified version of Mayer's contribution; write
* msp(a)ucsd.edu for the original code.  Kudos to Mayer for a fine piece
* of work.  -msp
*/

END QUOTE

The license for Pure Data may be found at
<http://www-crca.ucsd.edu/~msp/Software/LICENSE.txt>;, and it is indeed a 3
clause BSD type license.

However, it is not included in the source distribution of autotalent, and the
ownership of mayer_fft.c seems less than clear to me.  Can autotalent be
packaged as is?  If it can be packaged, do we need to include the Pure Data
license file as a second source file in order to comply with that license? Or
does "If it were up to me, I'd say go do whatever you want with it" mean we
can do whatever we want with it?

Also, the patents in question appear to be U.S. Patents 4646256 and 4878187,
which, if I understand correctly, have expired.

Thanks for giving this your attention.

David Cornette

[Tom Callaway]

On 05/14/2010 03:27 PM, David Cornette wrote:
>  However, it is not included in the source distribution of autotalent,
> and the
>  ownership of mayer_fft.c seems less than clear to me.  Can autotalent be
>  packaged as is?  If it can be packaged, do we need to include the Pure Data
>  license file as a second source file in order to comply with that license? Or
>  does "If it were up to me, I'd say go do whatever you want with it" mean
> we
>  can do whatever we want with it? 
I think that autotalent is fine to be packaged as is, with

License: GPLv2+ and BSD

I would strongly recommend that you include the Pure Data license file,
and, if upstream is alive, ask them to do so (and to properly note the
license in the mayer_fft* files).

~spot

[Orcan Ogetbil]

On Fri, May 14, 2010 at 4:24 PM, Tom "spot" Callaway wrote:
>  On 05/14/2010 03:27 PM, David Cornette wrote:
> > However, it is not included in the source distribution of autotalent, and the
> > ownership of mayer_fft.c seems less than clear to me.  Can autotalent be
> > packaged as is?  If it can be packaged, do we need to include the Pure Data
> > license file as a second source file in order to comply with that license? Or
> > does "If it were up to me, I'd say go do whatever you want with it"
> mean we
> > can do whatever we want with it?
>
>  I think that autotalent is fine to be packaged as is, with
>
>  License: GPLv2+ and BSD
>
>  I would strongly recommend that you include the Pure Data license file,
>  and, if upstream is alive, ask them to do so (and to properly note the
>  license in the mayer_fft* files).
>  
Hi, I have some general questions. What you say above is in direct
conflict with the review guideline in [1] which says:
"If (and only if) the source package includes the text of the
license(s) in its own file, then that file, containing the text of the
license(s) for the package must be included in %doc."

My take on this guideline is, we shall put the license file to the
%doc that is in the source tarball (necessity) only if there exist one
(sufficiency). Thus we are not supposed to include external license
files in packages.

The questions are, am I interpreting the guideline wrong? Or are we
making an exception for this package? If yes, what grants an exception
and is this documented?

Orcan

[1] https://fedoraproject.org/wiki/Packaging:ReviewGuidelines

[Tom Callaway]

On 05/24/2010 09:22 AM, Orcan Ogetbil wrote:
>  On Fri, May 14, 2010 at 4:24 PM, Tom "spot" Callaway
> wrote:
> > On 05/14/2010 03:27 PM, David Cornette wrote:
> >> However, it is not included in the source distribution of autotalent, and the
> >> ownership of mayer_fft.c seems less than clear to me.  Can autotalent be
> >> packaged as is?  If it can be packaged, do we need to include the Pure Data
> >> license file as a second source file in order to comply with that license? Or
> >> does "If it were up to me, I'd say go do whatever you want with it"
> mean we
> >> can do whatever we want with it?
> >
> > I think that autotalent is fine to be packaged as is, with
> >
> > License: GPLv2+ and BSD
> >
> > I would strongly recommend that you include the Pure Data license file,
> > and, if upstream is alive, ask them to do so (and to properly note the
> > license in the mayer_fft* files).
> >
>  
>  Hi, I have some general questions. What you say above is in direct
>  conflict with the review guideline in [1] which says:
>  "If (and only if) the source package includes the text of the
>  license(s) in its own file, then that file, containing the text of the
>  license(s) for the package must be included in %doc." 
Not really, think of it like this:

If (and only if) the source package includes the text of the license(s)
in its own file, then that file, containing in the text of the
license(s) for the package _MUST_ be included in %doc.

If it does not, and you have sufficient evidence of licensing for a
source file, _and_ that source file does not contain a copy of the
license text within itself _and_ the license requires that the license
text be distributed (for example, BSD does this), you MUST include a
copy of that license text. (You should try to get upstream to fix these
sorts of situations, as technically, they are also in violation of such
licensing terms.)

In all other situations, you are not required to include separate copies
of license texts with the Fedora package, however, maintainers may do so
at their own discretion.

>  The questions are, am I interpreting the guideline wrong? Or are we
>  making an exception for this package? If yes, what grants an exception
>  and is this documented? 
This is somewhat of an exception, because the BSD license text requires
it. Normally, we wouldn't hit this, because the source file would
contain the BSD license text within itself, and we wouldn't need to
include a separate copy, but in this case, the BSD license doesn't
appear anywhere within the autotalent source tarball, but we know
mayer_fft.c is BSD.

If the above wording is clear, I'll add it to the Licensing#FAQ.

~spot