Michael Eriksson
A Swede in Germany
Home » Humans » Thinking | About me Impressum Contact Sitemap

Putting claims to a test

A splendid example of how putting too much value on the “obvious” can lead one astray, and the value of putting claims, hypotheses, etc., to an actual test:

At E6 a discussion ensued about why keyboards were referred to as having 102 (105, other relevant number) keys. With one exception, we were all convinced that this number was vastly exaggerated, and I remember speculating that it counted many keys twice (with and without shift, e.g. “a” and “A”). The one exception said something about actually having counted the keys, but no-one actually listened to him—after all the claim was ridiculous, 28 letters, 10 digits, various special characters, etc. makes a total of possibly 60–70 keys.

Well, if one of us had bothered to count the keys on the keyboards immediately in front of us... Doing so with the keyboard I am using right now, I find:

  1. Top row: Escape, 12 function keys, three keys I have never used. Total 16.

  2. Numpad: 10 digits, num-lock, “/”, “*”, “+”, “-”, “,” and enter. Total 17.

  3. Arrow keys: 4.

  4. “Edit-block”: Page-up/down, begining/end of line, delete/insert. Total 6.

  5. Main area: 30 letters (the physical layout is German, although I use the US layout logically), 10 digits, 2 each of shift/alt/control, caps-lock, Windows key, menu key, tab, delete, enter, 8 special characters. Total 60.

This gives a grand-total of: 103.

Notably, this is not even an unusually populated keyboard: Many others include half a dozen to a full dozen “multi-media keys” and power-management keys, a second Windows key, and/or a separate toggle for switching the function keys from their default function to multi-media functions. Notebooks often contain yet other special keys (but usually do without e.g. a separate num-pad). Values above 110 cannot be that uncommon, even on an English “Wintel” keyboard—other plattforms and locales could conceivably reach even higher values. (At least historically, there has been a limit of 128 keys that can send separate codes; thus, going above 128 would rarely make sense.)

Looking into why this odd cognitive error occured, I suspect that the sheer multitude of keys not corresponding directly to a character (including function keys, shift, alt, Windows, arrow keys, ...) is too easy to overlook. The duplicated keys on the num-pad probably worsen the problem. Notably, a keyboard stripped down to the absolute minimum might well fit below 60 keys, certainly below 70.