My recent reading of Google collections makes me think that the right way to do a cache is with their MapMaker and computableHash method. Very nicely done.
On Sun, Sep 6, 2009 at 11:55 PM, Sean Owen <sro...@gmail.com> wrote: > My pet theory about caches is the caller should only need one > operation -- get(K key). The cache is instantiated with an object that > can load a value for a key. Given that it can do all its work. > Otherwise every caller has to have knowledge about how to manage the > cache and how to load values for keys. I am not sure, maybe this is > partly what you are getting at, in which case this approach solves > some of those issues? > -- Ted Dunning, CTO DeepDyve
