I have in columns past discussed my belief that many aspects of the open-source software movement are poorly managed and often work against the goals of the movement. In an earlier column, for example, I discussed in detail the problems with the basic structure of open-source licenses.
As a lot are crafted alike, they have defects in common, which show up in the frequent efforts by projects to use other licenses (such as those borrowed from the Creative Commons, one-off variants of existing licenses, or placing software directly in the public domain, which is, effectively, an impossibility in the U.S.).
The problem behind the license issue and behind most of the other aspects I am concerned with derives from the leadership of the open-source community. Putting aside the famous peculiarities that form a defining part of Richard Stallman’s persona, the rest of the “leaders” of the movement—to the extent that the open-source bazaar has leaders—have two fundamental traits in common: They view open source as needing protection, and they view open source as innately superior to other forms of software. Both views are dangerous illusions, in my opinion. Let’s take them in order.
The view that open source needs protection from those who would subvert it runs through all aspects of open source. Most prominently, it shows up in the licensing. The Open Source Initiative tried to trademark “open source” (a request rejected by the U.S. patent office, which handles trademarks), and it has certainly tried hard to define open source in terms of its own sanctioned licenses.
These licenses are unabashedly designed to prevent companies from releasing mixed open- and closed-source products. The fear is that companies will get the benefits of the open-source moniker without in fact making all the source code freely available. It is hard to know what exactly the perceived benefits of the open-source moniker might be, as for years, in fact, the label has been a bar to acceptance in businesses. (As I discuss later, this is finally changing.)
Open source does not need protection. Companies have found ways to circumvent the prohibition against mixed open and closed source (by releasing different versions: community editions and then closed-source higher-end editions), so protecting the purity of open source is no longer needed.
The second problem, that of a sense of the innate superiority of open source, was on view earlier this week. The Apache Software Foundation (ASF), which hosts several high-visibility projects, called Oracle to task for not letting it use validation tests as it wants. The ASF threatened to resign from the JCP if Oracle did not comply. Putting aside for a moment that almost no one outside of the ASF would care whether they were on the JCP, the choice to threaten Oracle is peculiar and disturbing. Why should the ASF get for free what other companies have paid handsomely for? The rationale derives from the prevalent view that open source is inherently better than commercial products, surely purer, and therefore deserving of special privileges.
This sense of special entitlement is odious to me. Most open-source projects of any size are developed by commercial entities, so they’re not volunteer efforts in the same way non-profit humanitarian groups are. Open source is simply a different model of commercial software development. Nothing else. This “nothing else” dimension is finally reaping benefits: Businesses are shedding their suspicion of open source, and they increasingly consider it as just another entrant in the mix of solutions they evaluate.
To continue this trend, which is crucial to the long-term success of open source, groups like the ASF have to abandon the idea that they are special and uniquely deserving. They’re not. The ASF is almost entirely supported by payments from commercial entities. At one time, the ASF was actually developing software, such as the Apache Web server and the Apache Portable Runtime. It’s been a long time since those days, and the foundation is no longer a development community as such, but rather a hosting site for contributed projects, supported by corporate entities and staffed in part by volunteers. (I discussed its future relevance in June.)
The ASF’s ultimatum to Oracle is a combination of both sentiments I view as wrong-headed: the misguided perceptions that open source needs to be defended (by the ASF), and that open source is innately deserving. Oracle does not help open source by giving in to this pressure. If the ASF will not meet the requirements set forth by Oracle for the validation tests it wants, then the group should do what every other commercial vendor would have to do in the same situation: Figure out how to work without them.
Andrew Binstock is the principal analyst at Pacific Data Works. Read his blog at binstock.blogspot.com.