Two technical reports have been released which provide a detailed comparison of various missing data procedures. In fact, the details may have obscured general philosophies and concepts. The purpose of this working paper is to discuss these philosophies and concepts. After two reports the time seems appropriate to reflect upon the trends in the work which is already done and to outline the future work which is needed. Thus, this paper has a more general, discursive and often subjective tone than the two previous reports, but this tone is needed in order to give an overview without digressing into statistical technicalities.