Writing is a key part of using a computer. The standard keyboard layout
(Qwerty/Qwertz) is less than ideal under certain criteria. In this thesis we
develop a program which can search for an optimal character layout for a
given physical keyboard model based on samples of text that are expected
to be written on it. The resulting layout has many advantages over a Qwerty/Qwertz layout such as: typing speed, lowering the number of writing
errors, lessening the strain on hand muscles. We achieve that by using a
genetic algorithm. The genetic algorithm simulates the use of the keyboard
on the given text for its fitness function. We focus on searching for a layout for typing in Slovene, however the program is capable of finding layouts
for any given language as well as specialized languages as say programming.
Using the program, we find two new keyboard layouts. We compare them
to Qwerty/Qwertz and some other established alternatives using certain statistical methods. We also run a short comparative study between them and
Qwerty/Qwertz.
|
