A recent study in Nature (PMID 16355180) showed that entries for 42 science-related articles in the Wikipedia are almost as reliable as entries on similar topics in the Encyclopedia Brittanica . This is a small sample, which is limited to science articles. But larger studies in multiple disciplines may validate this conclusion. If so, this would be empirical proof that open source/open access reference products are not inferior to traditional sources.
Wikipedia allows anyone to edit and re-edit entries, without requiring expertise in advance. All you need is an Internet connection and an interest in a given topic, and you're ready to write.
It is important for open access advocates to recognize that the Wikipedia model is not flawless. For controversial topics, writers on different sides of an issue sometimes edit and re-edit pieces ruthlessly. Wikipedia is taking measures to create "stable" articles that cannot be easily updated, and is also seeking out expert writers in various fields.
Even as Wikipedia matures, its central concept of building an openly accessible, democratically managed resource will remain the same. Open access advocates should see the Nature study (which is not OA itself, of course) as good news.