[conspire] Short permissive software licences

Rick Moen rick at linuxmafia.com
Tue Sep 29 17:39:52 PDT 2009


So, at the end of 2007, software author Daniel J. Bernstein got tired of
people ragging on him about his generous-but-proprietary non-license for
most of his software including Qmail and djbdns -- and came up with the
one option tailor-made to cause almost as many problems:  He has
declared that most of his works, which he stopped maintaining in the
1990s, are "public domain".

Since then, four different coders have started producing updated
versions of djbdns (and related software such as tcpserver and
ucspi-tcp).  Some of these guys consult with me on -- hey, guess what?
-- licensing problems resulting from Dan's 2007 chump move.

One of them asked me about a big patch from a contributor in Hungary:


 From: contributor in Hungary
 Date: Thu, Aug 27, 2009 at 5:34 PM
 Subject: Re: tcpserver-limits-2006-01-26.diff
 To: coder
                                                                                
 In my country jurisdiction it is not possible for author to legaly 
 release work to public domain (except by dying and staying dead for 70
 years), but I hearby grant everybody licence to use those tcpserver
 patches for any and all purposes without any obligations to me, and
 relicence and release combined work as they wish without need to
 contact me.

 so it should be as good as public domain as I can grant - eg. I don't
 reserve any rights on them.

 have fun,
 [contributor]


The coder sent that over to me with the question "Any suggestions?"
I just now replied:



 Date: Tue, 29 Sep 2009 17:25:59 -0700
 From: Rick Moen <rick at linuxmafia.com>
 To: coder
 Subject: Re: Public Domain Suck Part XVI

Quoting [coder]:

> Any suggestions?

Hi, [coder].  Thanks for asking.

My understanding of the law is that an unconditional disclaimer 
of interest, such as "I hereby grant everybody licence this 
work for any and all purposes without any obligations" is 
the maximal possible licensing, and to my knowledge should be
satisfactory to anyone.


If I were seeking to release software under permissive terms, I
would put a notice right in the source headers, like this:


# Copyright (C) 2009, Rick Moen
# Author grants any person permission to copy, modify, merge,
# publish, distribute, sublicense, and/or sell this software 
# and associated documentation files without obligations.

Above is a cut-down version of the MIT License pruned to minimal
length.  It omits for length purposes a disclaimer of warranty
that would be much wiser to include, e.g.,

# Copyright (C) 2009, Rick Moen
# Owner grants any person permission to copy, modify, merge,
# publish, distribute, sublicense, and/or sell this software
# and associated documentation files subject to the condition
# that there is no warranty of any kind, expressed or implied,
# including but not limited to warranties of merchantibility,
# fitness for a particular purpose, and noninfringement.

The MIT License is discussed at: 
http://en.wikipedia.org/wiki/MIT_License

Anyway, the above shows what is, to the best of my understanding, the
best way to release software under a particular licence:  First, you
have a line that states you copyright interest and the year it came
into existence.  The form I used is the standard way of saying that.
Then, below that, you state your terms as copyright owner.  Ideally,
this is right at the top of the covered file(s), or in a "LICENSE" or 
"COPYRIGHT" file in the same tarball.


You're probably asking me whether you should accept [contributor]'s
licensing conditions.  I'm sorry to have to hedge my bets, but I have to
be careful about not violating "unauthorized practice of law" statutes,
since only an attorney licensed to practice in your state is permitted
to give you advice on specific legal situations you're in.

Thus my advice (above) is necessarily stated in general terms rather
than ones tailored to your situation.  Sorry about that!

I think [contributor] is _trying_ to create a maximally permissive
"academic" licence.  I really wish software authors wouldn't do that,
because of the risk of accidentally getting something wrong.  You might
consider asking [contributor] if he's willing to place his work under
the MIT License, as that is a short (~20 line) permissive licence that
everyone knows.


I once tried to create the shortest possible permissive licence that
(I think) covers all required points.  This is what I came up with:

# Copyright (C) 2009 Owner Name. Do whatever you want with this work.

This is the variant with warranty disclaimer:

# Copyright (C) 2009 Owner Name. Do whatever you want with this work,
# provided you agree that there is no warranty of any kind.


I'm cutting corners somewhat to save lines.  For example, the reason the
MIT License lists "copy, modify, merge, publish, distribute, sublicense,
and/or sell" explicitly is to make the inclusion of all those
otherwise-reserved rights crystal clear.  By saying instead "whatever
you want", I'm trusting that judges will read my intention correctly.
Also, the reason warranty disclaimers typically list all of the 
types of implied warranties explicitly is to make really clear that,
really, none of them is granted.  (You may have noticed that warranty 
disclaimers are most often in all capital letters.  That's because 
the Uniform Commercial Code section 2-316, adopted by all US states
to govern commercial sales, requires that warranty disclaimers be
"conspicuous".  See:
http://www.law.cornell.edu/ucc/2/article2.htm#s2-316 )






More information about the conspire mailing list