Python 2.0b1 is released!

Grant Griffin g2 at
Tue Sep 12 09:32:06 CEST 2000

Paul Wright wrote:
> >    From my understanding, CNRI is located in Virginia and so it makes sense
> >for them to want the license to be interpeted under the laws of the state
> >they're in.  I really don't think that this is a plot to use UCITA to carry
> >out some underhanded scheme.  I would be more worried if CNRI wanted their
> >license to be interpeted under some state that they have no connection with.
> I believe the point here is that, according to what I've read, the UCITA
> allows the licence to change retrospectively: that is, CNRI can apply
> the current 1.6 licence now but in Virginia, they can legally substitute
> another licence at a later date which is binding on people who *already*
> have the software.
> This from <>
> "UCITA will allow the publishers to impose the most outrageous
> restrictions on you. They could change the license retroactively at any
> time, and force you to delete the material if you don't accept the
> change. They could even prohibit you from describing what you see as
> flaws in the material."
> So, you're not only trusting CNRI now, but in perpetuity (or at least
> until UCITA is suceeded by something else).

I can't help but point out that in believing that interpretation, you
are trusting the FSF.  I don't really know whether or not their
intrepretation on this issue is fair or accurate, but like any
self-respecting iconoclasts, they definitely have a bias here that one
should be aware of.

IANAL, but as I understand it, depending on who you ask, software
publishers might have _already_ had the ability to change the license
retroactively.  For example, in the case of the very generous license of
Python 1.5.2, CNRI is now telling us that it wasn't really a license, so
that's why Python 1.6 needs something much different than what each and
every user of 1.5.2 _thought_ was it's license.  (Which begs the
question: What exactly _was_ the license for 1.5.2?!)

But have you ever noticed that software licenses seem to be largely
academic anyway?  Here's what I have observed:

1) Every license of every kind has a "disclaimer of warranty".  Even
Microsoft and GNU can agree on this point.  Nobody who makes software
(or anything else) wants to get sued.  But I haven't heard of anybody
sueing for defective software.  I don't know if this is a result of the
disclaimer or not.  Honestly, it probably just isn't feasible to sue
people for something that costs maybe $500 or less.  But specifically in
the case of free/open software it is unlikely that one would lose any
kind of "implied warranty" lawsuit: if the software was both free and
open, the plaintiff would seemingly have a hard time explaining why he
thought you owed him money when it didn't work as he intended, and when
he was welcome and invited to fix it.

2) For private users, all software license seem to boil down to "the
honor system".  There really isn't much to make you follow the terms, if
you don't want to.  For example, once you have software, it's rather
hard for anybody to take it away from you; if you don't want to delete
the software, it's unlikly that the UCITA folks will send dragoons
around to "rough you up".  As another example, if you include a little
GPL'ed code in a program you distribute as an executable, there isn't
much to force you to follow the terms of the GPL; it's pretty unlikely
that anybody will discover this grave slight to the Noble Cause of
Freedom, and it's even _more_ unlikely that anybody will come after you
for doing it. 

3) For commercial users, it makes business sense to play by the rules,
whatever those may be.  However, the rules are open to wide
interpretation, and except for the very most grave and obvious breaches
(e.g.  The Big Corporation installs Microsoft Office on all its
computers worldwide without paying for it), it is basically infeasible
for software publishers to enforce the terms of their licenses.  So what
I have seen in Big Corporations I have worked for in the past is some
sort of half-hearted attempt at license compliance, which really doesn't
assure that each and every copy in use was paid for.

But these practical observations aside, people tend to see software
licenses not just as legal entities but as moral imperatives (for some
strange reason.)  But since the law provides considerable gray area, I
encourage all software users to stay "legal" but interpret any software
license liberally, and in their own interest; in a practical sense,
there seems to be little trouble one can get into for doing so.

>From this perspective, in the present case of the CNRI license, one can
see very little practical downside to marrying Python with GPL'ed code,
if one has that desire.  Therefore, even though Richard Stallman may
wake up in the middle of the night screaming over this, the rest of us
should respond to the CNRI license's use of Virginia (and UCITA, by
inference) in accordance with the following large, friendly letters:


calm-ly y'rs,


Grant R. Griffin                                       g2 at
Publisher of dspGuru                 
Iowegian International Corporation

More information about the Python-list mailing list