A recent study found that out of eight open source mobile platforms, Android ranks at the bottom in terms of openness. Photo: Jim Merithew/Wired.com

By Ryan Paul, Ars Technica

Market research firm VisionMobile has published a report that evaluates the openness of eight major open source software projects. The study—which was partly funded by the European Union—focuses largely on open governance, inclusiveness, transparency, and ease of access to source code. To quantify relative openness, the researchers established criteria and a numerical rating system with points.

The projects that VisionMobile analyzed include Android, Eclipse, the Linux kernel, MeeGo, Firefox, Qt, Symbian (based on the governance model of the Symbian Foundation prior to the the platform’s transition back to a closed model), and WebKit. They ranked these projects in an “open governance index” based on the percentage of points that they received. Google’s Android mobile operating system ranked the lowest, with only 23 percent. The Eclipse integrated development environment ranked the highest, with 84 percent. Android was the only project in the study that scored less than 58 percent.

Android’s low ranking in the index came as no surprise to us. As we have written on severaloccasions in the past, Google’s mobile platform falls far below the standard of openness that the search giant promised when Android initially launched. The VisionMobile report identifies some of the key problems with Android’s governance model, including Google’s “unilateral Android project decision-making processes” and “closed contributions process model.”

“Visibility to the roadmap is limited, as there is no Android roadmap publicly available. In fact, development of the Android private branch and the roadmap is controlled by Google, with little input from external parties or the Open Handset Alliance members,” the report says. “When launched, the Open Handset Alliance served the purpose of a public industry endorsement for Android. Today, however, the OHA serves little purpose besides a stamp of approval for OHA members; there is no formal legal entity, no communication processes for members nor frequent member meetings.”

The lack of transparency in the Android compliance program is also identified as one of the weak areas in Android governance. The report says that “Google tightly controls the Android platform and its derivatives” by using its unilateral control over the full Android compliance criteria, which are “undocumented and somewhat capricious.”

Although the actual Android compatibility definition is public, the specific parameters of the compliance testing are a closely guarded secret. Skyhook is currently pursuing litigation against Google for allegedly doctoring the compatibility testing parameters for anticompetitive purposes.

The report quotes an internal Google e-mail (made public as a result of the ongoing Skyhook litigation) in which Google’s Android compatibility chief Dan Morrill described the company’s compliance testing practices as “using compatibility as a club to make [OEMs] do things we want.”

Browser openness

Android was, of course, just one of the eight projects discussed in the report. The study’s scores for Firefox and WebKit are also intriguing. WebKit had a slightly higher score (68 percent) than Firefox (65 percent).

The main areas where Mozilla lost points related to the lack of public data about project contributors, the size of the developer base, and the number of commits from community members. Mozilla’s impressive new contributor metrics dashboard, which was announced in April but is currently undergoing a security audit prior to public launch, will likely address those shortcomings.

Eclipse was identified as the most open project in the study. In particular, VisionMobile lauds Eclipse’s Project Dash and general commitment to transparency. The organization’s vendor neutrality and well-defined governance structure are highlighted as positive governance attributes.

A look at the criteria

The full score tables and numerical criteria are included in the report. A close look at the scoring turned up some interesting discoveries. The criteria is a bit subjective in places, but generally looks good. There were, however, a few aspects that might be debatable.

Oddly, VisionMobile’s criteria assigns a one point bonus for mandatory copyright assignment, a practice that is not generally regarded as a hallmark of good open governance (for some good background on the issues with copyright assignment, see the commentary by Dave Neary and Michael Meeks).

The VisionMobile report itself acknowledges that none of the projects included in the study require copyright assignment and that copyright assignment is probably unnecessary. (It’s more common for contributor agreements to stipulate a perpetual copyright license rather than outright assignment.)

Another aspect of the criteria that might be controversial is the scoring for licenses. The study awards a higher score for using a permissive license than using a copyleft license. Although permissive licenses increase the flexibility of downstream code use, they aren’t necessarily conducive to more open governance.

Money matters

The report contends that greater openness generally leads to greater success among open source software projects. At the same time, it acknowledges that Android’s popularity contradicts that conclusion. The Android “paradox” is discussed at length in the report; it suggests that “Google’s financial muscle” and engineering resources, rather than openness, have driven Android’s success.

The full text of the VisionMobile report is available under a Creative Commons license and can be downloaded from the firm’s website.