Abstract
This paper proposes a method for incorporation of the scientific notion of a system of measurement units into programming languages. A method proposed in a previous paper on this topic is first discussed and found to be seriously flawed. We find also that the earlier method must rely on run-time checking in order to allow certain desirable language features, whereas the method proposed here is completely implemented at compile-time, and permits improved program self-documentation, and is clearly more compatible with the spirit of Pascal, the language chosen for this exposition. Further, the actions expected of the compiler are very modest indeed, and a suggested implementation method is given.

This publication has 0 references indexed in Scilit: