Hi Can someone confirm my belief that the following code in Products/StandardCacheManagers/RAMCacheManager.py is incorrect (upon which I will raise a collector issue) try: from cStringIO import dumps except: from pickle import dumps I believe it should be try: from cPickle import dumps The outcome of this error is that the results from every object that is associated with a ram cachemanager instance is pickled to see if it can be cached. And by not using the optimised version, which will always happen because "dumps" is part of cPickle not cStringIO, we always get the "slow" version. Thanks Tim P.S. Doesn't pickling the cacheable results seem to be a very expensive way to determine the suitability of caching the object?