This survey must begin by drawing attention to competing accounts of the subject. First, there is WOLFGANG RUPPERT'S book [55], This advertises itself as being about separately continuous semigroups but in fact contains a lot of material about semigroups with multiplication continuous on one side only. Then there is NEIL HINDMAN'S survey [24]. This, even more improbably for an article which contains a substantial amount on compact right topological semigroups, claims to be about ultrafilters and Ramsey theory. Finally there is the book by JOHN BERGLUND, HUGO JUNGHENN and PAUL MILNES -not the familiar volume of 1978 in the Lecture Notes series, but the entirely new book which will appear in 1989. This offers a superb elegant account of the whole theory of topological semigroups and their compactifications, not as deep as RUPPERT'S masterful summary of his special area, but providing a real headache for anyone attempting to produce an alternative over-view of the field. I shall, of course, try to say a little about all aspects of the subject, but in order to avoid too much overlap I shall concentrate on those areas which are more familiar to me and about which I may, at this moment, have more information than other authors.Any reviewer of this field faces another, much more irritating problem: terminology. I shall be dealing with semigroups in which the multiplication (s,t) i-> st is continuous in one of the variables. Obviously it does not matter which is chosen, and different authors have made different choices. If we decide on continuity in s, we might describe this as 'left-continuity' because s is the left-hand variable. But we might equally observe that it is the operation of translation on the right by the element t which is continuous and use the term 'right-continuity'. Again, individual authors have followed their own inclinations. What is worse, if you find a writer's terminology unpalatable, you cannot simply ask your word processor to go through a paper interchanging 'left' and 'right', because everyone agrees what (for example) a left ideal is. It is obviously time that a decision was made on a uniform terminology. In spite of the consequences for some of my own preferences, I hereby declare that BERGLUND, JUNGHENN and MILNES [6] will become the standard reference work in this area; I shall follow their terminology and notation and I recommend all other writers to do the same.There is one further aspect of notation about which an innocent reader should be warned. In this subject, commutative semigroups can be dense in non-commutative Unauthenticated Download Date | 6/15/16 5:45 PM 198 John S. Pymones. For this reason the practice has grown up of sometimes using the expression 5 +1 to denote a non-commutative product of s and t, and we shall do this when convenient. An addition sign does not necessarily imply commutativity.