http://www.dinkla.net/fp/thesis.html
The implementation of geometric algorithms is often considered as difficult. This is due to the following three problems: inexact arithmetik, degenerated inputs and the correct implementation of complicated parts. The traditional method of software development is made responsible by researchers for the gap between theory and practice of geometric algorithms. We note, that the first two problems are caused by the nature of geometric problems and the last two problems by the methods of software development and the programming language used.
In the thesis is shown, that complex geometric algorithms can be implemented short, modular und with a high degree of abstractness in the functional programming language Haskell. Therefore functional programming possess the potential to reduce the two last mentioned problems.
