Darwin: A matter of privacy?

Apple's Darwin license reveals a difference between "free software" and "open source."

COMMENTARY--It's been a week of mixed news on various fronts. The death of Douglas Adams (author of The Hitchhiker's Guide to the Galaxy and one of my favorite authors) and the death watch over Eazel (more on that in a future column) have been pretty depressing. But the week also saw some good news that, in part, resolves a controversial complaint I made in a column two weeks ago. Unfortunately, the good news also opened another can of worms.

In the column two weeks ago, I angered scores of Apple and BSD fans by calling Apple's claim of "embracing of open source" this year's Big Lie, and by saying that the company had not given back anything substantive to the open source community.

One of my main complaints was that Apple opened up Darwin, the company's Mach-based OS, under a license that wasn't recognized anywhere as open source. When I wrote the original column, the Apple Public Source License (APSL) was explicitly classified as non-free by the Free Software Foundation (FSF), and it hadn't been recognized as an open source license by the group that coined the term, the Open Source Initiative (OSI).

Since neither the OSI nor the FSF had characterized the APSL as free or open source, I was accurately able to state that Apple's contributions couldn't be counted as open source. But that situation changed recently because of a simple vote.

Last week, mere days after my column on Apple appeared, the OSI voted to endorse the APSL as an acceptable open source license, and it may even show up on OSI's list of acceptable licenses by the time you read this. Of course, that the vote happened last week refutes the claims of respondents to my column who cited a two-year-old press release indicating the OSI's endorsement of Apple's actions as proof that the APSL was open source all along.

If the OSI had liked what Apple was doing back in 1999, why didn't it endorse the license then? The answer was in the version of the license. What the OSI didn't say in the press release was that the version of the APSL it was looking at in 1999 contained clauses that prevented it from actually meeting the Open Source Definition, OSI's published criteria for open source licenses. Apple addressed those issues when it released version 1.2 of the APSL early this year, and then the OSI was able to approve it.

With this vote, the OSI nullified my main complaint against Apple, and the company can indeed be said to be working with the open source crowd. The vote doesn't address my other suggestions about how Apple could help the open source community; but as respondents to my column were quick to point out, Apple isn't obligated to do any of that. We'll have to agree to disagree about whether Apple has been a sufficiently good neighbor. But one thing is clear--Apple's claim that it has embraced open source can no longer be considered an untruth.

Having said that, the resolution of one issue appears to have brought another to the forefront: In the current version of the APSL we now have a major license--perhaps the first--that is explicitly accepted as "open source" by the OSI but is explicitly rejected as "free software" by the FSF. In the discrepancy we can see the first substantive issue illustrating that the difference between "open source" and "free software" isn't purely philosophical.

That issue is privacy.

In the explanation of why the APSL is non-free, the FSF notes that many previous objections to the APSL have been dealt with by the current version, but that one deal-breaker remains. Clause 2.2 of the APSL requires you to publish (and register with Apple) not only code that you change and redistribute, but code that you change and only use internally.

The FSF says that this is an invasion of your privacy, and that you should have the right to keep private any modifications that you don't redistribute. The definition of "open source," however, doesn't say anything about privacy.

"The registration requirement is a real (though in our opinion minor) privacy problem," said OSI head Eric Raymond. "But the Open Source Definition was not written to address privacy problems." Raymond added that there is no current or foreseeable movement within the OSI to include privacy provisions, but "if there arose a general feeling that privacy issues needed to be within the OSI's scope, we would respond."

The upshot is that, for better or worse, we now have a specific issue by which to differentiate the OSI vision of "open source" from the FSF vision of "free software." The terms are no longer interchangeable. While the FSF has always maintained a distinction between the two terms, until now the difference has been mainly of philosophy and intent. While both the FSF and OSI have endorsed the BSD, Apache, Perl, and other licenses, the APSL privacy issue appears to be the first situation where a license can be considered "open" but not "free."

Now that Apple has its open source license, much (but not all) of my complaint against the company is gone. Yet, in the process, the gap between the OSI and the FSF has widened.

But enough about Apple, the FSF, and the politics of software licensing. I'm going to have another listen to my recording of the original radio version of The Hitchhiker's Guide to the Galaxy. Have a pleasant voyage, Douglas.

Does the difference between "open source" and "free software" have any effect on you? Let me know in the TalkBack below.


You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All