When a man sits down to write a letter to himself he is seldom troubled by semantics, but this letter is an exception, since the topic, in my opinion, is entirely semantic and hinges on what we mean by the word ‘definition.’ At the last two Society meetings there have been lengthy discussions aimed at a formal definition of operations research, and a number of potential authors threaten to submit long articles on this subject for publication. Operations Research, ISSN 0030-364X, was published as Journal of the Operations Research Society of America from 1952 to 1955 under ISSN 0096-3984.