Last changed 13 Jan 2019 ............... Length about 900 words (6000 bytes).
(Document started on 13 Jan 2019.) This is a WWW document maintained by Steve Draper, installed at You may copy it. How to refer to it.

Web site logical path: [] [~steve] [educ] [this page]

Dijkstra's paper: my footnotes

By Steve Draper,   Department of Psychology,   University of Glasgow.

These are just my notes, mostly from my first read through. I have no real idea at the moment, whether these are of low, medium, or high interest to others.

Really, I also want to locate each note back in a section;
I'd like to only have to write the id nmb once, not twice.
And I'd like an extra section on collected key phrases/arguments summing up a
section when possible.
[1] section B2d   (para.4) This argument amounts to an argument for shallow, not deep, learning; avoiding links with existing knowledge. This is to some extent anti-constructivist. Or conversely, we could seen him as an unusual voice pointing out the advantages in some cases of shallow learning, which is not generally admitted in educational writing.

[2] section A3   Scale. He's wrong. "Deep time" as dealt with by geologists and then biologists like Darwin have established it, and dealt with it. We are now pretty certain that the age of our sun and earth is about 4.7 * 109; and the universe about 3 times that age. If you work at it, you can get a grip on these scales, but you do have to work at it. On the other hand, you also have to work to get a grip on much smaller numbers such as the distance in miles (or kilometers) to places you know; dates such as 1860 mean nothing until you have drilled yourself in a handful of dates in that century.

Also: see Paul Nurse's view that there needs to be, and is coming to be, and fifth great theory in biology to tackle the huge complexity of a single cell: about 105 chemical reactions proceeding simultaneously in each cell.

[3] section B2c   His anti-anthropomorphic campaign. However a) "talk to" usually IS about communication b) Biology and imputing intention to evolution or to habits: they often discuss it and generally don't think it's too big a trap.

[4] section A3   Operational thinking about behavour: but you should reason =about set.s But note Bertrand Russell's point that (in 1913 when he wrote this) the highest most developed science tends NOT to be about causation, but about invariants and relations. You can "predict" where a planet was long ago as accurately as where it will be in the future.

[5] section A3   I think he's wrong in some ways, but we'll never have a decent theory and practice of CCSE until we either accept and implement his ideas OR have better arguments about what is wrong with them.

[6] section A2   He says teaching is bad by always presenting things as easy, familiar, and so unexciting. And this is also a problem because it tells the student that it well be easy to learn and no need to think much about it: leading to shallow learning. But he doesn't recognise that this is a permanent dilemma in education: it's just as bad to the do the opposite, and put them off trying by emphasising the novelty and difficulty. In fact we should recognise what is new; and what is an extension of the old.

[7] section A3   Bayard: talk about what you haven't read; operate around what you couldn't build yourself nor understand.

[8] section A3   Testing: only test for errors already known and common. BUT testing for unheard of errors (testing to discover the unknown unknowns) is necessary, done by all sensible engineers; but essential but untheorised. And so not faced.

[9] section A3   Digital designs that violate continuity: stairs, cogwheels. The piano not the cello.


[91] section A3   The two novelties he identifies are:
  • Scale (109 not 103);
  • Digital not continuous. Not continuous change in input leading to a continuous change in output. This is not completely true. A program processing a range of numbers is more likely to have errors at the ends of the range than in the middle.

    N.B. integers are digital vs. real (continuous) numbers in maths

  • [92] section A4   These novelties are radical because they have to be tackled differently.

    [93] section A3   These novelties are radical because they have to be tackled differently.

    [94] section B2   Educational problems: because of using fundamentally faulty metaphors; because of using old methods for new material.

    [95] section B1   Scientific consequences: Computing is using human symbol manipulation to create an abstract symbol manipulation (the code); which when supplied to a computer, executes mechanical symbol manipluations. Deal with ALL elements of a set by ignoring them and dealing with the definition of the set.

    Web site logical path: [] [~steve] [educ] [this page]
    [Top of this page]