I just pushed up some experimental, alpha quality code for asynchronously persisting Ruby objects to disk at my GitHub repository. Again, this is very early, alpha quality code with minimal, incomplete testing. I took the idea that Redis and other Key-Value stores use to provide fast reads/writes while providing persistence. I plan to allow for more customization with different types of triggers (like Redis allows), but for now it simply synchronizes the data on a regular interval.

Allows for persisting single objects or a class of objects (though there are some conflict resolution problems with multiple objects pointing to the same file).