An anomaly in the proof of a theorem characterizing nonreducible flow graphs is pointed out. A proof is given which corrects for the anomaly and is fundamentally simpler than the original proof.

A Computer Science Department is no longer a novelty in our universities. On the other hand, the place of computer science in the university, the subject matter to be taught, and the way in which the teaching is done vary widely from one institution to another and with time at any single institution. My concern here is with an aspect of curriculum and of facilities which seems to have been inadequately appreciated. The subject matter is the massive array of programs constituting the software system through which the computer is used. My contention is that we have not understood the problem of teaching software systems and are not properly equipped to teach much more than “Some Aspects of How Our Supplier's System Works.”

A syntax for a language to express algorithms in a way which facilitates proofs of equivalence between algorithms as well as translation into computer programs is given and briefly justified.

Two distinct algorithms for computing the generalised factorial function defined by Cooper (1966) are written in a FORTRAN-like language and then, without use of the principle of recursion induction (McCarthy, 1962) are proved equivalent.

A basic technique in the design of a nuclear power reactor is to prescribe the dimensions and composition of a simplified model and to calculate the corresponding neutron flux and power distributions.Comparison of calculations for different models and of particular calculations to experimental results guides the physicist to a reasonable prescription.Until recently, the usual model had spherical or cylindrical symmetry.We have found it feasible to extend the methods to include a class of reactors in which there is variation of composition with respect to two space-dimensions.In this paper, we summarize some of our experience with these calculations.