Subject: Re: Can OSI specify that public domain is open source?
From: Chad Perrin <perrin@apotheon.com>
Date: Wed, 7 Sep 2011 15:47:50 -0600
Wed, 7 Sep 2011 15:47:50 -0600
On Wed, Sep 07, 2011 at 05:50:00PM -0400, Karl Fogel wrote:
> Chad Perrin <perrin@apotheon.com> writes:
> >
> > Derivative, involving additions under other licenses -- yes.
> > Modifications to the original -- no.  BSD and MIT licenses still
> > require inclusion of the license in modified versions.  
> 
> That's slightly different.  They require that because whatever original
> code remains is still under that original license.  But the presence of
> that original header doesn't mean that your new code is automatically
> under the same license.  You may choose to "reuse" the header, because
> you like that license.  But you might include another copyright notice
> that describes the different copyright on your changes.

Are you somehow claiming that removing the license from the file is
legally kosher, then?  Considering it would directly contradict the text
of the license, I find that a bit difficult to believe.  About the only
way I can think of to make your scenario work is to distribute your
changes as a diff.


> >
> > The public domain example requires re-dedication to the public
> > domain, as I understand it, every time someone modifies it.  The
> > inheritance characteristics are notably different.  I just don't
> > believe that disqualifies it as open source software, in the spirit
> > of the OSD.
> 
> Publishing new changes to (say) BSD code also requires
> "re-notification", it's just that the most common way to achieve the
> desired result is to leave the existing license header intact.

It requires leaving the notice where you found it (attached to the work)
while you modify other parts of the work.  That's not the same as
changing a typo and having to *add* some kind of license, dedication, or
waiver notice.  My point is that no specific inclusion is required by
law for modifications of public domain works; you must make a conscious
decision to do so, because the license does not apply to the modified
work (there being no license).  Meanwhile, one needs to go to significant
effort to actually route around the license in the case of modifications
to a work covered by the license in its entirety.


> 
> PD doesn't require re-dedication to PD, by the way, but if it's to
> remain open source it *does* require some statement about the new
> changes, that ensures receivers that the new code is also open source.
> That statement could be a supplemental PD dedication, if such a thing
> were possible to do reliably (apparently doubtful), or it could be a
> standard license (easier).

That was my point about the public domain.  Um . . . thanks for restating
it?

-- 
Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]


["application/pgp-signature" not shown]