I'm adding my thoughts and perspective related to a timely and insightful post (thanks for pointing that out, Molly) about the relevance of open source and community over at the SnapLogic blog site.
The post touched on points about open source communities (the read/write web ratio as it applies to active/non-active members, small numbers of actual source code committers, etc.) as well as the role of active development. However, the heart of the piece was put forth as this:
Open source actually does not matter at all when it comes to harnessing community contributions. It’s not the open source, it’s the open APIs!
Since this sports a high probability of being misinterpreted (as flamebait) if read too quickly or taken out of context, I will go on record as saying that the providing of open API's does indeed encourage a bulk of value-added contributions to any community. When access to any variety of underlying architecture (whether Facebook, Google Earth or Eclipse) is granted through an API kept transparent/standards-based enough to be dubbed 'open,' the fact that the architecture is featured as open source itself becomes irrelevant to those who desire to contribute through the API. Thereafter, the API is the point-of-entry with everything else losing a degree of importance.
Yet, at the point where an API has been published and has found significant use by a developer community (and the open source/proprietary element is a side issue) the aforementioned underlying architecture must have reached an according level of stability while having demonstrated reasonably attractive functional capability. Which, ironically, is a state that, featuring an open source software development model serves as an accelerator in reaching. In other words, open source affords the shortest path to the stage where the matter of closed or open is basically a non-issue.
Open source also lends an air of credibility to the technology underlying newly introduced API's, which is actually important as developers don't just pick up any old API and start banging on it. The transparency afforded through open source can inspire a sense of trust in those who may be unsure one way or the other. One such example of leveraging the open source model until critical mass is reached and things become API driven, is SugarCRM. The company initially utilized the cost proposition afforded by the open source model to capture select pockets of the CRM marketplace and mature the product until it was an established player with a proven business value proposition that also just happened to be an open source vendor as well. As of today, SugarCRM is a cost-effective, powerful and flexible CRM system as well as a solid CRM platform. Most importantly, a capable web services API allows the SugarCRM platform to be easily integrated within web service friendly IT ecosystems. It's really not a stretch to assume that SugarCRM as a closed source option would have found reaching its current stage a Herculean effort, if it ever came close at all.
Outside of Web platforms, like Amazon or EBay which have value rooted in governance by the companies behind them, reaching critical mass is more than half the battle and open source makes that journey quite a bit more feasible. As a result I find open source, a noted accelerator, difficult to categorize as simply a useful feature. To me it represents a structural building block for boosting platforms which are exposed through the very same open API's mentioned earlier.