The Python 1.6 License Explained

[Pat McCann]

"Andrew Dalke" <dalke at acm.org> writes:

>   I used the old CWI license as the basis for biopython.org's license,
> although with a few name changes as appropriate.  Given the answers
> in #1-3 (copied below, though trimmed to make a bit smaller), does that
> mean the biopython.org code has the same legal problems, that is,
> not being published, not a legally binding license and able to be
> revoked?

You should get a lawyer to confirm all this, but here's my opinion:

Not being published is not necessarily a problem.  It makes a difference
in certain circumstances, but you'll have to study copyright law fairly
deeply to figure it out.  It won't be much of an issue as long as people
who accept your license can know that the copyright owner has offered
the license in a legal sense and that the license covers the software.

>   If so, should I convince everyone (about 5 people now) to switch
> over to the new BSD copyright - the one without the advertising clause?
> Though it seems also to lack clauses concerning these "problems."

Without such clauses some people will wonder if the license is good
enough.  Many think it is and that is not revocable, but it's just
opinion.  Many lawyers wouldn't buy it, but many have.
 
>   Or what about adding statements like "this is a legally binding
> document which cannot be revoked except by material breach of one or
> more terms" and "presence of licensed source code and any accompanying
> text constitutes publishing."

Statements of (possibly false) fact like that don't make the fact true.
There are certain things that make a license agreement binding or make
an activity publishing.  Look these things up in an Internet law
dictionary for hints to their meanings.  You can start at google.com

http://www4.law.cornell.edu/uscode/17/, esp. Section 101 is good reading.