Well, idea for my class came in the thread "Estimating amount of memory need for cache" (more than week ago). Of course the way of measuring memory size in the Florin's link is best possible, but only when nothing else is happening in the system. My way could be possibly used in the working system, but is dependant on the implementation if the JVM. I don't know if it is reasonable to use my way at all (depending on the JVM implementation??). What do you think (Clinton?) Michal Malecki > I was sure it can't be that easy :(
But looking at Florin's link IMHO it becomes very easy. Comparing memory usage before and after object creation is a nice estimate or do I miss something? NIcholoz Koka Kiknadze
