Licensing Python code under the Python license

"Martin v. Löwis" martin at v.loewis.de
Sat Mar 12 11:32:39 EST 2005


Ville Vainio wrote:
>     Daniel> Thanks for the advice.  I'll probably go with either the
>     Daniel> BSD license, or possibly the LGPL.  But I'm leaning
>     Daniel> towards the BSD since it fits on the screen...
> 
> Isn't MIT license even shorter and simpler? A while ago some Debian
> guys were speculating whether even BSD license is "free enough" to
> include in Debian...

I encourage anybody to read Larry Rosen's book on this matter,

http://www.phptr.com/bookstore/product.asp?isbn=0131487876&rl=1

Larry describes what he likes and dislikes about each of the licenses
from a legal point of view - giving insights you cannot possibly get
without law school.

For example, the BSD license reads

"Redistribution and use in source and binary forms, with or without 
modification, are permitted provided that the following conditions are 
met: [...]"

Compare this to the rights that a copyright holder has, e.g. from

http://www.whatiscopyright.org/

"exclusive right to reproduce, prepare derivative works, distribute, 
perform and display the work publicly."

Strictly speaking, the BSD license gives non of these rights to the
licensee. The right to redistribute is probably *meant* to include
the right to reproduce - or is it meant to allow distribution only the
very copy that you received yourself (so you have no copy after 
distribution). The right to use is not one that copyright law has
control over, so what does it mean that the license gives you that
right? (*)

What about the right to prepare derivative works? Most likely,
the license is *meant* to give this right also, since you are
permitted to redistribute modifications (but then, perhaps only
modifications of the original author?). And so on.

Larry argues that a license should be legally meaningful, and
legally clear - or else there is little point in formulating
a license in the first place. If the license is formulated
ambiguously, in the case of doubt, courts will have to interpret
them. While courts are capable of producing such an interpretation,
they sometimes do so in a surprising manner (*).

I've been picking on the BSD license because I can remember
the complaints Larry has about its text.

Regards,
Martin

(*) If you are curious: Larry argues that, while the permission to
use is meaningless in copyright law, it is meaningful in patent
law. To use something, you need a license for all patents that
would otherwise prevent you from using it. So the permission to
use *could* be interpreted to be a patent license. However, most
likely, the authors of the license did not intend it to be a
patent license - so what the right to use is remains unclear,
until courts rule on this aspect.



More information about the Python-list mailing list