Subject: Re: Can OSI specify that public domain is open source?
From: Karl Fogel <kfogel@red-bean.com>
Date: Wed, 07 Sep 2011 17:06:43 -0400

John Cowan <cowan@mercury.ccil.org> writes:
>You modify another work named "bar" that has a public-domain notice.
>Let's assume for the moment that the work truly is in the public domain;
>perhaps it is written by Barack Obama in the scope of his employment.
>You publish your modified version.  Is the work still in the public
>domain?  Arguably no!  It is now a proprietary work (though without a
>copyright notice, so you will find it hard to sue).
>
>That's what makes public-domain notices really bad.  People assume that
>they can be treated like open-source works, but they cannot, not without
>highly unexpected consequences.

No, this would be just as true if the original work had been under a BSD
or MIT license.  You still couldn't assume that someone's random
derivative work is under the same license as the original, unless they
explicitly attached headers indicating that.

Now, this header attachment happens "automatigically" if the changes are
in the same file, but that's sort of a logistical accident: it's not
really about copyright law, it's just about how people who would have
forgotten to do something they meant to do are saved because the natural
process of editing files leads them to unintentionally make a claim that
they might have made intentionally if only they'd thought of it.

If their changes involve adding new files, then even that won't happen.

And suppose they make changes for internal use, where they (or their
company) actually *didn't* intend for the changes to be open source.
Then that code gets released accidentally somehow.  Now many people
might take the derivative work to be open source, when it's really
not open source.

So I think this reasoning is a red herring.  PD is complicated, but this
particular objection doesn't apply to it any more than it applies to
licensed open source works.

-K