In this new post to his blog Kevin Schroeder suggests that there's something even better than doing the typical caching inline (request, write to cache) - pre-caching.

I just had an epiphany. I've talked about pre-caching content before and the benefits thereof before. But this is the first time I realized not only that there are benefits, but that doing it is BETTER than caching inline. Let me sum up... no, there is to much. Let me explain.

He gives an example of how a typical application might cache - when it finds a "miss" for the data it's trying to pull. A simple cache is easy, but what happens if it uses a configuration value that could change (like the username/password in his second example). Pre-caching would eliminate the risk since the setting would be known to be valid when the cache is generated.