Since Makoto Nagao first pronounced the ``machine translation by analogy principle'' in 1984 until now the so-called paradigm of ``Example-Based Machine Translation'' (EBMT) has emerged and developed the original idea. This talk gives an up-to-date overview of recent developments in the field partially based on the contributions of a workshop on EBMT last September. It will be shown that EBMT subsumes a number of heterogeneous approaches using a variety of methods and different sources of background knowledge.