The keybowl: an ergonomically designed document processing device
Assets '96 Proceedings of the second annual ACM conference on Assistive technologies
The metropolis keyboard - an exploration of quantitative techniques for virtual keyboard design
UIST '00 Proceedings of the 13th annual ACM symposium on User interface software and technology
Ant colony system: a cooperative learning approach to the traveling salesman problem
IEEE Transactions on Evolutionary Computation
| Hi-index | 0.00 |
Keyboard design is a combinatorial problem of perennial interest to those who work with computers. An evolutionary algorithm can search a space of key arrangements for one that is easy to use. Its genotypes represent candidate keyboards, and its fitness function measures the difficulty of typing a specified body of text. Here, a genetic algorithm seeks to arrange the twenty-six letter and six punctuation marks (four of the latter in two pairs) on three rows of ten keys each. It encodes keyboards as permutations of key positions. Its evaluation function considers the relative loads on the hands, consecutive keystrokes by one hand or finger, and the relative difficulties of using individual fingers as it simulates typing a block of text. In tests using a concatenated body of text and its three components (a collection of message logs, Moby Dick, and the King James Bible), the GA consistently evolved keyboards with markedly better evaluations than either the QWERTY or Dvorak keyboards, though the GA's keyboards differed depending on the text it used in its evaluation function.