Diatribe

BOARD OF COMPUTER STUDIES TEACHING SUB-COMMITTEE A DIATRIBE
Alan Creak, l4 March 1979.

1. Monotribe on Assembly languages.
	My emphatic espousal of the cause of Assembly language teaching is, equally emphatically, not to be construed as support for Assembly language itself; all the Assembly languages which I know or have met are abominations. Neither am I concerned that the students in the Mathematics 191 class should know a great deal about any particular version of computer hardware. But I believe that it is very important that a student should have a reasonably accurate mental picture of the sort of thing that happens in a computer when a subroutine is entered, or an array element is selected, or a parameter is passed - or, for that matter, when an instruction is executed. For all its defects, Assembly language speaks simply and clearly about such things, and I know of nothing else which does the same thing so directly.
		REMARK If a student does so understand, he will understand the difference between name and value parameters, why there are problems in implementing recursive procedures ( or aren't, if he starts with a stack machine ), why Fortran arrays are laid out as they are, and so on. To do so, though, he must know about storage, addressing, indexing, and indirection: these are hard to get across without a clear picture of how a computer works, and a language in which to describe it. KRAMER
	The reason for my insistence on this sort of low-level acquaintance with computing is that I cannot imagine any other way by which a student can come to UNDERSTAND computers. If the object of the Mathematics 191 is to teach students to use computers without understanding them, then many of my objections vanish, and I urge the immediate establishment of the Department of Necromancy, and the transfer thereto of Maths 191. ( No doubt there are several other courses in the University which could also profitably be transferred. ) If, on the other hand, the course is to introduce people to computing and computers in an intellectually respectable way, then I can see no alternative to starting at the bottom. Further, I think that the bottom in question is the way in which an instruction is executed; there is obviously room for discussion on where lines should be drawn, but the idea of a machine following a set of instructions seems to be a major point of difficulty for many people, and much becomes clear once this hurdle is past.
		REMARK This remains true even for - maybe especially for - the student who does not intend to do any more computing. We don't need a general public which can write Fortran programmes - we need a greater understanding of what computers can do, what they can't do, and why. KRAMER
	As a final point, I readily accept that Assembly language may not be the only way to get across the material which I see as essential - but I can't see a realistic alternative. In fact, of course, it isn't really very difficult. Maybe we do it the wrong way in Mathematics 191, but I suggest that the main troubles there are the limited time available for the topic, and the compulsion to give a reasonably ambitious assignment to appease the continuous assessment ogre.

2. Monotribe on Mathematics and computing.
	First, I am not too concerned what happens to the courses in Computational Mathematics; they have little or no contact with, or bearing on, the courses which I see as central to a Computing Department, and will finish up in the Department of Computing or Mathematics ( or Necromancy ? ) as is best suited to administrative convenience. I am more concerned that the fuss over computational mathematics should not lead to neglect of courses in other mathematical topics which are much more relevant to the core Computing courses; I have in mind such topics as theory of automata, combinatorial properties of data structures, complexity of computations, linguistic theory, and so on. ( I may have got the names wrong - I'm not as well informed as I would like to be on these topics myself. ) These all have a very direct bearing on the subjects taught in the present stage 2 courses, but there seems to have been no suggestion that a Department of Computing would offer anything at all along these lines. At present, some of the Board's courses include some introductory material on the theoretical aspects of their subjects, but - especially as more and more non-mathematicians enter the courses - these courses do not provide a good context for formal mathematical treatments. Computing is in the happy position that most of its important points can be communicated by example, or pictorially; the mathematics is not necessary for understanding. The mathematics is nevertheless an important study in its own right, and I believe that a department could very profitably offer a "practical" course, which most people would take, and a parallel "mathematical" course which would delight the mathematical enthusiasts. Any such development would, of course, be very difficult to bring about if the "mathematical" topics remained ( where some of them at least now are ) in the Mathematics Department - I am not questioning the goodwill of all concerned, but recognising the practical problems of getting people in different departments close enough together to achieve a profitable collaboration. ( We know a lot about that problem in Computer Studies. ) I believe that there would be much to be gained from teaching the mathematics of computation in the Computing Department, and I think it should be discussed before it's too late.

CONCERNED :	My real concern was that it seemed to me that the eminences of the Board of Computer Studies did not always give the impression of knowing what they were talking about. I didn't suppose that they were stupid, but rather suspected that some had a limited, or partial, view of the topic. I was certainly unimpressed by some of the Board's pronouncements.
CONTINUOUS ASSESSMENT :	A bad thing. If students can't learn the stuff without being forced to do pointless assignments, they should fail. Optional assignments might be constructive.
GENERAL PUBLIC :	The original 191 course was not directed at people intending to make a career out of computing, or even necessarily to do much of it; it was for people who would like to know enough about computing to make sense of it. That still seems like a good idea. Do we do it ?
MONOTRIBE :	Yes, I know it's bad etymology; it's a joke. Ha ha.
REMARK ... KRAMER :	The "REMARK ... KRAMER" pair was a sort of pun on the current fashion for constructing bracketing pairs in just that way in programming languages - for example, CASE ... ESAC, IF ... FI . I assumed that anyone likely to be a member of the Board of Computer Studies would be familiar with current computing practice. This was a mistake; I have no numbers, but several weren't. I can only hope that they understood the points I was trying to make.
UNDERSTAND :	See above. I thought that understanding was the idea. I think that our earliest groups of students wanted to understand, and did. I am a lot less confident about some of those I met in the stage 3 Operating Systems course around 1998; their grasp of how computers worked was sometimes tenuous in the extreme ( though no doubt they'd done all the compulsory assignments ).

An attempt to inject a drop of reality

Notes.