I'm not planning right now to make any changes to the license text
because I think that would only make things worse.
But I'll respond anyway:
On Sat, Jun 27, 2009 at 9:34 AM, Cosmin Truta<cosmin@...> wrote:
> Glenn Randers-Pehrson wrote:
>>> Now that this discussion has started, I would like to propose a fix to
>>> libpng's ever-increasing license.
>> "ever-increasing"? Except for addition of contributing authors
> ... which consists of 4 lines of extra text, added every time a new
> contributing author comes around, not counting the contributing
> authors themselves.
> The zlib license is small and easy to read. It has 14 lines, not
> counting the blank lines. The libpng license, on the other hand, is a
> 65-line "mutant" license. When I use libpng version X, why would I
> care about the terms and conditions pertaining libpng versions a, b, c
> and d, where a<b<c<d<X?
As the user, you don't care. As the redistributer, We should be
interested in obeying the conditions of the license under which
we received libpng.
> This may seem an insignificant issue to some people, but I have
> learned that some organizations actually require a legal review
> whenever a license text of the software they're using has changed.
> Changing the "copyright bits" (as pointed to by Bob), by adding one or
> more new contributing authors, should be fine. But in the libpng
> license, the overly verbose text that accompanies each set of new
> contributing authors, is not just part of the copyright bits. It also
> contains terms and conditions.
Only once it contained additional terms and conditions.
>> there is only one
>> addition, that seemed at the time to be necessary in the face of
>> UCITA (http://en.wikipedia.org/wiki/UCITA). But in the end the UCITA
>> proposal was withdrawn and only implemented by Maryland (my state)
>> and Virginia.
> If you are referring to the addition to "against infringement", even
> that thing was unnecessary. I remember Adam pointed this out in a
> discussion on the cexcept list: when you say "no warranty", that
> should mean zero warranties, not one warranty (of non-infringement...)
> Besides, none of the other important relevant OSS licenses out there
> have undergone changes like that.
UCITA says otherwise. The text of the additional paragraph is mandated
by UCITA. That is important to me because I live in Maryland which is
one of the two adopters of UCITA.
All of the other OSS licenses are vulnerable, or would have been if
UCITA had caught on.
>> It might be appropriate at this time to eliminate that clause, leaving
>> us with the original libpng license plus additional contributing authors.
>> But that still wouldn't make it the "zlib/libpng license" as defined
>> by OSI or the "License of Zlib" as defined by FSF, so I'm not sure
>> that deleting the clause would serve any useful purpose.
> I'm ok with that. After all, it's a half-sentence.
Eliminating it does not actually shorten the license. It merely adds
a line or two saying that the paragraph below is no longer
effective, or something along that lines.
It may be that I could really eliminate it, since it was added on
That would of course hurt our comment/code ratio. #:-)
>> It isn't just concurrence of Guy and Andreas that we'd need. All
>> of John Bowler, Kevin Bracey, Sam Bushell, Simon-Pierre Cadieux,
>> Andreas Dilger, Magnus Holmgren, Tom Lane, Dave Martindale,
>> Glenn Randers-Pehrson, Eric S. Raymond, Greg Roelofs, Willem van Schaik,
>> Guy Eric Schalnat, Paul Schmidt, Tom Tanner, Cosmin Truta, Gilles Vollant,
>> Tim Wegner, and Group 42 would need to concur in a change to the license
>> (How many of us are still here?).
> Are they all copyright holders? I thought that the only copyright
> holders are Guy (and Group 42, Inc.), Andreas, and you. The
> contributing authors are said to disclaim all warranties, express or
> implied, etc., etc., but they aren't copyright holders, so they should
> have no say on this.
That's probably correct. But without Group 42's approval (I already have
Guy's and Andreas' somewhere) I can't do anything really useful.
> Glenn, if what you're saying were true, then any open-source project
> with tons of contributors would be a licensing disaster!
Yes. I have already been contacted several times by legal departments
wanting signatures of all of the contributing authors. I told them to
try to live with the license that we provided, without additional
>> Back to my original question, what do we call our license? From the
>> discussion, it seems to me that it's the "libpng license", despite
>> what OSI says. And zlib uses the "zlib/libpng license".
> Yes, I think that "the libpng license" is an appropriate name.
> Ideally, we should let OSI know that what they're referring to as "the
> zlib/libpng license" really is "the zlib license".
I think they already know (or knew) that.