Can Microsoft Truly Issue 'Open Source' Licenses?
Yesterday, publisher Tim O'Reilly broke the news that at his company's own open source convention in Portland, Oregon, Microsoft General Manager of Platform Strategy Bill Hilf is planning to -- if it hasn't already -- submit its existing Shared Source Licenses to the Open Source Initiative, for certification as true "Open Source Licenses." The OSI is the designated caretaker of the legal definition of "open source."
But the question may rightly be asked: Is this a genuine move by Microsoft to enroll its Microsoft Permissive License (Ms-PL) as an official open source license that the community can recognize, or is this more of a symbolic act?
In a blog post yesterday, Microsoft Director of Source Programs Jon Rosenberg offered this explanation: "Today, we reached another milestone with the decision to submit our open licenses to the OSI approval process, which, if the licenses are approved, should give the community additional confidence that the code we're sharing is truly Open Source. I believe that the same voices that have been calling for Microsoft products to better interoperate with open source products would voice their approval should the Open Source Initiative itself open up to more of the IT industry."
But that explanation was embedded -- or perhaps more accurately described, spliced -- into the midst of an essay that meandered through a multitude of different topics without a common thread or segue. While it discussed a sort of phase change in Microsoft's open source history between being a "trailblazer" and being a "road-builder," it later diverged down a road less traveled, touching upon the sticky subject of whether the OSI best represents the open source community if its membership is selected by a board and not the community itself.
Traffic laws and drivers' licenses are very necessary things, Rosenberg for some reason conceded, though it's always nice to have lawmaking bodies whose membership is determined by the people at large. At the time when a reader might wonder just where is this going, Rosenberg made the OSI license announcement excerpted above.
And then, without segue, he rendered the following: "So what about the flip side of the OSI becoming a membership organization? Could they really be voted out of existence or rendered ineffective? It doesn't seem likely to me. Participation in the OSI and adherence to OSI licensing guidelines and Open Source definitions is entirely voluntary. If it isn't serving the best interests of the community, the community will go elsewhere."
Rosenberg provided a link to the OSI Web site, specifically to a blog post from last June 15 from its president, Michael Tiemann, entitled, "Designing a New OSI." In it, he writes:
"The board of the Open Source Initiative has largely concluded that we have reached a point of organizational and contextual maturity. Namely, that open source has been defined, and a relatively large constituency of people have accepted that definition. So far, so good. What we have not done, however, is to make the OSI representative of that constituency. Yes, our board members have strong credentials as open source software developers, entrepreneurs, advocates, researchers, etc. But we cannot really claim that we are truly representative of the community, nor that we can truly speak for the community, other than the fact that each of us considers ourselves to be a small (very small) part of the community. There are others who identify themselves as members of the open source community (just as we do), who strongly believe that they better know how to protect and grow the open source movement, which includes greatly relaxing our standards for interpreting the OSD and allowing a great many more licenses to be approved."
It's unknown just which company Tiemann was referring to in that latter sentence, though perhaps Rosenberg's strangely embedded link to it may give you a pretty strong clue.
Under its current rules, OSI openly invites developers to submit their licenses for approval, which would mean they meet the formal definition as OSI perceives it. Based on a quick read of those rules, it would appear the Ms-PL would meet the basic tests of "open-sourceness" spelled out by OSI. Distributed works must be freely distributed, and contain source code. Derivatives must be allowed, so long as they are freely distributed as well. No discrimination of licensees must take place, and the license cannot restrict itself to any specific technologies.
By those rules, Microsoft's current alternate shared source license, called the Limited Permissive License (Ms-LPL) would fail, because it restricts redistributions to Windows platforms only.
But in an open source version of "Catch-22," there's another rule which could prevent the basic Ms-PL from being certified as well. It's not a rule of "open-sourceness," but rather has to do with the motive behind the submission.
Specifically, the process asks submitters to tell OSI which existing OSI license is most like the one being submitted, but then explain the need for the change. In other words, the submission must explain why the licensor can't just use the existing OSI license. In a splendid bit of irony, since Ms-PL is so short and, quite likely, so respectful of the current open source definitions and processes without adding anything to them, it could be rejected for those reasons.
A rejection would be a noteworthy event which the press would probably cover and analyze, with one of the issues inevitably becoming, who made the decision about who decides what's open source and what isn't...and suddenly Jon Rosenberg's blog entry doesn't seem so cryptic.
Yesterday's news came in the midst of Microsoft's inauguration of a marketing Web page to support its open source efforts, in addition to its Port 25 blog from its Open Source Software Lab. The marketing page contains direct links to Port 25 blog entries. In his announcement yesterday, Bill Hilf said the new page "clearly outlines Microsoft's position on OSS by providing specific information about Microsoft, the OSS community, and the interaction between the two."