Truth and Clarity in Arithmetic

Being further items relating to an essay in Computer, vol.36, no.2, 2003 February, pp.108,106-7.  The items are Errata, further explanation of the operational calculator, Harold Thimbleby's study of calculators, further consideration of editorial keys, further discussion of basic operations, and some other relevant URLs.

Errata

On p.108, middle column, second paragraph, the statement "numeracy in developing countries shows steady decline" whereas I actually meant to write "developed countries."  I have no data on numeracy in developing countries, but there is evidence (see below) for such decline in some developed countries, including my own (see for instance Kevin Donnelly's Australia needs to think again about its unis or section 4.1 of Changes in Academic Work (pdf))

Ralph Place has taken issue with my use of accuracy in "the calculator's accuracy greatly exceeds the slide rule's," where I had meant "the accuracy I can achieve with a calculator greatly exceeds what I can achieve with a slide rule."  See Sphere for more on accuracy and precision of slide rules, this introduction for more on slide rules, here for details of Otis King's cylindrical slide rule, and, closely related to slide rules, nomograms are described in the excellent Blood, Dirt, and Nomograms by Thomas L Hankins and published in Isis, vol.90, 1999, pp.50-80.  By way of contrast to these analogue calculators, the little remembered Curta calculators competed with slide rules in their heyday, and still have a fan club.  Incidentally, IBM's first electronic digital computer, the 701, was commonly referred to as a calculator.


The Operational Calculator

Several e-mails have led me to believe that my explanation of how one aspect of clarity is to be achieved, namely, how the operational keys are to be directly linked to the operation marked on the key, was inadequate.

  1. After any operation the (minumum) four lines of the display will typically hold, reading from the top, the first operand of the operation, the second operand of the operation (prefixed by the operation symbol), the result of the previous operation (which is to be the first operand of the next operation), and a blank line, the bottom line.
  2. If a dyadic operation is to be carried out next, the user keys the value of the second operand into the bottom line.
  3. The user then keys in any prefix required for the next operation.
  4. Finally the user taps the operation key, whereupon the requested operation is directly and immediately carried out. 
  5. The display is replenished in the fashion described in the first step above.
  6. If the user wants to carry out a monadic operation then the second step above is not carried out (the second operand field is left blank), and the resulting content of the lines is slightly different.

Thimbleby's Study of Calculators

Harold Thimbleby has made a thorough and persistent study of calculators, as the study cited in the essay shows.  But he has gone further and designed an excellent basic calculator which works on an algebraic or declarative basis rather than an arithmetic or operational one (A True Calculator, A New Calculator and Why it is Necessary).

The declarative calculator maintains an equation, the operational calculator carries out arithmetic operations.  The arithmetic of the declarative calculator is inexact, that of the operational calculator is exact.  It seems to me that the operational calculator is ideally suited to support the kind of arithmetic I learnt in primary school, whereas the declarative calculator is ideally suited to the algebra and trigonometry I learnt in secondary school.


Editorial Keys

The essay mentions editorial keys for the basic calculator.  Such keys are not strictly necessary for a basic calculator, but four keys, marked say with up- down- left- and right-pointing arrows, are sufficient to provide a capability much greater than that of the four M keys of the typical commodity calculator.

The following description assumes that the basic calculator can be in either of edit state or history state.

In edit state there is a cursor active in the line where the next second operand is to be entered.  In edit state, the cursor is either in insert mode, in which case the cursor appears as a vertical line between character spaces, or in replace mode, in which case the cursor appears as an overlay of the current displayed character.  Keyed digits and signs either insert or replace at the cursor, depending on the mode.  Any operation key triggers an operation, after which the cursor is in insert mode.  The arrow keys have the following effect:
up arrow  to history state in scan mode, highlighting most recent number
down arrow  to history state in remove mode, highlighting most recent number
left arrow  move cursor one position left, deleting the character at the previous position if in delete mode
right arrow  move cursor one position right, deleting the character at the previous position if in delete mode
left & right arrows  toggle between insert and replace modes

In history state one of the previous operand values is highlighted in about the middle of the display area.  Displayed values cannot be edited in history state.  Actions in history state may be in scan mode, wherein the highlight is an underline and no alteration to the stored values is possible, or in remove mode, wherein the highlight is a video reversal and the highlighted value can be removed.  Keyed digits move the highlight the appropriate number of extra values (that is, one plus the value of the key) in the direction of the previous movement, deleting the previously highlighted value if in remove mode.  Any operation key (and prefixes may be keyed) causes the highlighted value to be copied to the edit line and the keyed operation to be carried out immediately. The arrow keys have the following effect:
up arrow  move the highlight one value up (back), removing the previously highlighted value if in remove mode
down arrow  move the highlight one value down (forward), removing the previously highlighted value if in remove mode
up & down arrows  toggle between scan and remove modes
left arrow  to edit state in replace mode, with the cursor to the extreme left
right arrow  to edit state in insert mode, with the cursor to the extreme right


Basic Operations

Some e-mailers have slated the calculator as I described it for having unfamiliar advanced mathematical operations, referring in particular to the count operation.  However, the count operation is both basic and arithmetic, and is neither advanced nor mathematical.

Counting is not only widely useful in everyday arithmetic, but is the basis for the long division that is, or was, taught in elementary schools.  Thus, if 7 is to be divided into 654321 then five counts of 7s must be carried out.  Note that this is associated with four or five modulus or remainder operations, the other operation that I associated with the octothorp key in my essay.

It is perhaps no coincidence that Snyder and Mitchell suggest in their paper Is Integer Arithmetic Fundamental to Mental Processing? that the similar operation of equi-partitioning is behind the remarkable arithmetical capabilities of many lightning calculators.

One of the reasons why counting is unfamiliar is that it is so basic that there is no primary symbol identified with it.  Even the vertical bar associated with the remainder operation in high school mathematics is not used in primary school.


Other Relevant URLs

  • John Allen Paulos, Innumeracy (this is probably the best known book on the subject, though somewhat old; a review with links)
  • Ned Rozell, Innumeracy (a website specific to innumeracy; here are recommended books)
  • Tim van Gelder, Numeracy (a new website collecting links; few as yet)
  • Iain Murray, Living with the consequences of innumeracy (some figures and insights from the Statistical Assessment Service)
  • BBC, University students "lack basic maths" 2002 September 27 (universities across the UK are having to offer undergraduates special tuition in mathematics to bring them up to scratch)
  • Mary Riddell, What is education for?, The Observer, 2002 September 29 (the handwringing about A-levels must not obscure the fact that schools are badly failing our children)
  • BBC, Value of classroom computers questioned 2002 October 28 (use of computers for teaching does not improve pupils' achievements and may even be harmful)
  • BBC, Pupils failing basic skills, 2002 December 4 (nearly half of Scotland's second year pupils fail to reach minimum levels in reading, writing and arithmetic skills)
  • Erica Cervini, Teens warned on calculators The Age, 2003 January 26 (teenagers who rely too much on calculators at school may miss out on an understanding of basic numeracy)
  • Brian Doig, Summing Up: Australian numeracy performances, practices, programs and possibilities, ACER, 2001 (no clear cut changes)
  • Failing maths and science, The Economist, 2003 April 17 (maths and acience have gone into a dangerous decline in Britain; background)
  • Maths "should be optional", BBC, 2003 April 21 (teenagers should be able to drop maths and study more practical numeracy skills, making the important distinction between mathematics and arithmetic/numeracy)
  • Mobile gadgets offer new lessons, BBC, 2003 May 28 (a European project is looking at how mobile phones and handheld computers can be used in education, for example to help build numeracy)
  • Making numbers add up, BBC, 26 July 2003 (with more than four in 10 people thought to be short in number skills, experts say it's time to change the way we think about numeracy)