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

Re: Migration of non-free packages to testing

On Wed, Nov 13, 2002 at 11:18:34PM -0800, Thomas Bushnell, BSG wrote:
> :)  I think you have to examine the specific case.  The point is that
> you can't evade the GPLs rules by saying "I refuse to distribute the C
> code, so that means the assembly code is the preferred *available*
> form."

But in that case, the C code does exist, so it's still "preferred among
the forms which actually exist".

In fact, I have *never* seen source code that would be my preferred
form in an ideal situation.  Usually the comments and/or the identifier
names suck, and it's often in the wrong language :-)

I also think that if an author *uses* a form for making modifications,
then that's a good first approximation of the preferred form for doing
so.  Who would know better than the author?

Of course, tricky cases can come up with code that so far no-one has
wanted or needed to modify.  ROMs for emulators can easily fall into
this category -- they are so old that their only use is in emulators,
and any change just makes them poorer at that task.  What is the
preferred form for making modifications if no-one prefers modifications?
The passive voice gets you into trouble there :)

I would say that a work is free if there are no artificial barriers
to making and distributing modifications to it.

Not necessarily related to all of the above: I have often wondered
about the status of revision histories.  If a project is maintained
via CVS, and the CVS repository is not open to the public, and only
exported snapshots (i.e. releases) are published, are those snapshots
really "source"?  Obviously the maintainers prefer to use the whole
repository when making modifications.  The logs and history in there,
the discarded and potential bits of code, as well as the alternate
branches and twigs of development, provide a richer context than is
available to anyone outside the maintainer group.  I think that
providing a snapshot stripped of its history is similar to providing
source code stripped of its comments.  In fact, a lot of things that
are kept as comments in some projects (ifdefed-out code, "it used to
be this way" remarks, "changed because of this" comments) are simply
deleted in projects where the maintainers rely on having a revision
history available.

Richard Braakman

Reply to: