The key concept in clustering objects (each with m attributes) of a data set is ________ which is typically represented (inversely) by ________ between them in an m-dimensional space?