Imagine my surprise today when I saw an email from an Intel lawyer in my OSI license-discuss queue...inquiring whether OSI would allow Intel to withdraw the Intel Open Source License, their vanity license!
Way to go, Intel!
I haven't written on license proliferation here yet, but now seems a perfect time to do it. Must point out that this is my $.02 and not a statement on behalf of OSI.
Recently some folks who are interested in open source have been griping about license proliferation and claiming that OSI (disclosure: I have been a board member of OSI since December 2001) has done a bad job of keeping the number of open source licenses low. Interesting to me that these claims should come up now and that they should seem to be coming primarily from corporate voices. (Yes, I realize I work for a corporation too, but generally I don't speak from that point of view :-)). It seems interesting to me (and I'm not the only one) because corporate vanity licenses that have caused the bulk of the problem.
Of course a low number of licenses was not originally an OSI goal. We were trying to get more businesses to participate! And participate they have. Unfortunately nearly *all* of the large organizations who wanted to participate also wanted a vanity license. To be fair, many of the vanity license generators just wanted to use the Mozilla Public License model, and to do that they had to re-author the license because the MPL has hard-coded references to Mozilla. And as long as the license has to be changed, why not customize ...get a license with the company or project name on it...add a couple of extra clauses that meet the business goals more exactly....or at least change it so the company's lawyers like the wording a little better?
Over the years all of us at OSI have participated in conversations with vanity license authors trying to convince them that they don't need to write that new license. They have nearly all been pretty sure that we "just don't understand their unique needs". I was recently involved in one such effort on the corporate side...the much-discussed CDDL. I refused to have anything to do with CDDL until Sun agreed to actually try to write a generic license which could be reused without reauthoring. It turns out this was not a trivial exercise, but I still think Sun did a good job drafting CDDL and that it will be seen as a positive contribution to ending license proliferation because it spawned the debate and forever raised the bar on what can be considered an acceptable new free and open source license.
As an aside, let me say that it took a long time for me to come to the realization that lawyers are a lot like programmers...except license text is their "code". Like programmers they are inclined to re-write to "make it theirs". It takes a leap to deeply realize that re-use is an open source value, that its actually offensive to open source people to suggest that "my license is better than yours because I wrote it".
So in the end we approved a number of vanity licenses because they complied with the OSD, knowing that the effect of all those vanity licenses would be not much license reuse (why use the Apple's license if you are IBM for instance) and not much true community around those vanity projects.
Vanity licenses are their own reward since they seldom result in healthy open source communities. The MPL was arguably the first vanity license as it was written by a corporate team for a specifc project and tried to split the difference between the copyleft and "copy center" concepts (although notice that the hallowed MIT/BSD and GPL licenses are all referential of specific organizational affiliations). In 1998 MPL was as far as Netscape was willing to go in their maiden foray into open source. But after a few years the Mozilla community came to realize that they were ready to go further into copyleft, and so they added the LGPL to their licensing strategy. This is a natural progression which I believe we can expect to see in other vanity projects.
In my opinion moves like today's surprise offer from Intel are the best way to end license proliferation...to have the original proliferators step up and reconsider. There are still many large organizations who are only just coming to open source and they will probably wish to initially use a license that takes the same middle-of-the-spectrum approach Mozilla first took more than six years ago, but hopefully after this debate and resulting actions OSI can now take, we'll have better luck convincing them to re-use one of the existing "template" licenses instead of authoring their own. What would be reallly cool is if the Mozilla.org folks got it together to create an improved and templatized MPL v2 so that everyone could agree on a definitive expression of most imitated open source licensing model.
Hi, Good work... Thanks
;)
Posted by: Waclaw | July 18, 2007 at 02:53 AM