The Economist's blog: maybe we should make our data free?
projectredstripe.com
projectredstripe.com
I think an interesting space, in particular, is those companies that distill large amounts of data into small relevant sets. Google is the first example that comes to mind, but search is not the only approach to this. Another example is YC's own LikeBetter, who sneakily collected lots of training data before launching this PAT thing. You can now figure exactly how your individuality personality compares to thousands of others, based on these millions of clicks they've receieved.
Are Google and LikeBetter making data free, or the opposite? There are trade secrets involved there, but we might not consider that "data". You can see interesting views and presentations derived from the data (search results, or personality tests), but you can never see all the data at once... nor would you want to.