Re: licenses in generated m4 files

[Alexandre Duret-Lutz]

>>> "Peter" == Peter Simons <address@hidden> writes:

 Peter> Bastiaan Veelo writes:
 >> [There] is no need to discuss this, because as has been
 >> pointed out before:

 >> "Every nontrivial file needs a license notice as well as
 >> the copyright notice. (Without a license notice giving
 >> permission to copy and change the file would make the file
 >> non-free.)"
 >> (http://www.gnu.org/prep/maintain/html_node/License-Notices.html#License-Notices)

 Peter> Well, my point is that the files (a) _are_ trivial because
 Peter> they are just automatically derived versions of some other
 Peter> file and that they (b) _do_ contain a license because there
 Peter> is an URL pointing towards the page stating it.

Not really.  The URL points towards a page that gives the
license of a macro that appears on this page.  That is not
necessary the same version you are using.  Also the macro could
have been removed from the archive, or the archive could have
been shut down, in which case you have lost the license.

Also, if XML is dropped, I don't think there is any point in
generating the M4 file.  (Documentation can be generated from
the M4 file, of course.)  I think stopping that would even make
the archive a lot simpler and useful:

  1. people use the same file that was submitted 
  2. you can resubmit a file that you have downloaded
  3. using the archive from CVS or from the tarball makes no
     difference (I have only been using CVS, so I never noticed
     that the macro on the website were generated until this was
     mentioned here, and I have never downloaded the tarball)
  4. the macro is self-contained and contains its license and
     documentation, in readable form (although less formated
     than one the web); so you don't need a browser and and an
     Internet connection to use it
  5. distributing the archive is easier

 Peter> For over 6 YEARS the Macro Archive has distributed macros
 Peter> without repeating the legal disclaimers in every single
 Peter> file, and not once has this bothered any of the users, but
 Peter> apparently NOW it really is a major problem.

I fail to understand why you react this way.  It is a major
problem only because you just called it so.  The age of the
error probably indicates that it is nothing but major, yet that
is hardly a reason for not fixing it.

[...]

 Peter> (1) How does the author choose which license his macro is
 Peter> under? How does the markup format specify this
 Peter> information? Who takes care of editing the existing
 Peter> macros to comply with this change (in case that will be
 Peter> necessary)?

Wouldn't allowing free text at the top of the macro do the trick?
That seems to be the place were one would put the license (keeping
the description of the macro near the macro).

dnl Copyright (C) 2004 Harry Burns
dnl
dnl Copying and distribution of this file, with or without modification,
dnl are permitted in any medium without royalty provided the copyright
dnl notice and this notice are preserved.
dnl
dnl @synopsis HB_CHECK_SALLY
dnl
dnl Define HAVE_SALLY if the Sally interface is required.
dnl this should be used to define conditional APIs as follows:
dnl   
dnl   #if !HAVE_SALLY
dnl     int order(int menu);
dnl   #else /* HAVE_SALLY */
dnl     int order(char* first_course,
dnl               char* first_course_exception,
dnl               char* main_course,
dnl               char* main_course_exception,
dnl               char* desert,
dnl               char* desert_exception,
dnl               ...);
dnl   #endif /* HAVE_SALLY */
dnl
AC_DEFUN([HB_CHECK_SALLY],
[

])

 Peter> (2) Which licenses do we accept?

Anything that is compatible with the license of the archive itself.

 Peter> (3) What does the generated m4 source look like for each
 Peter> respective license we support?

That one is easy: exactly the same as the submitted source :)
-- 
Alexandre Duret-Lutz