[erlang-questions] OT: typing distance

jm jeffm@REDACTED
Tue Jun 12 05:14:13 CEST 2007

As a matter of interest I decided to find out if a different keyboard
layout would make much difference to the physical effort involved in
programming as the main selling point of alternate keyboard layouts is
the reduced typing effort. Most of the examples only consider plain
English text and ignore the numbers and brackets which are used alot in
many programming languages, especially languages like lisp and erlang.
For example, consider these two calculations

Test 1

cat .../lib/erlang/lib/*/src/*.erl > erl_lib.txt

and feed the resultant file into keymilage from

and got these results

Dvorak 498mi 353' or approx 498*1.6+353*0.3/1000 = 796.91km

Qwerty 634mi 3522' or approx 634*1.6+3522*0.3/1000 = 1015.46km

wc erl_lib.txt
1002246 3258048 33462358 erl_lib.txt

so per character we have

Dvorak 796.91*1000000/33462358 = 23.8 mm
Qwerty 1015.46*1000000/33462358 = 30.3 mm

Hence, Dvorak in this test comes out the better of the two.

Test 2
There is another typing metric application at


which seems to have a problem with such large amount of text. So using

Distance: 1574m  Number row: 9.178%
Same hand: 23.76%   Top row: 25.64%
Same finger: 5.923% Home row: 58.27%
total keys: 86311   Bottom row: 6.896%
Finger: 5% 10% 12% 10% -- 11% 8% 13% 26%

Distance: 2175m   Number row: 14.93%
Same hand: 41.24%  Top row: 37.81%
Same finger: 9.098% Home row: 28.54%
Total keys: 86311   Bottom row: 18.70%
Finger: 4% 7% 18% 14% -- 12% 9% 13% 19%

In this test Dvorak is better in the distance metric again, left right
balance, and using the home row more. It is not so good on the right
little finger though; Note its high usage at 26% for Dvorak vs 19% for
Qwerty caused by the placement of the brackets.

There's really no reason to post this except others might be curious
about this as well.


More information about the erlang-questions mailing list