The Theory of Classification Part 20: Modular Checking of Classtypes

Abstract

A first-order type system has two things to commend it. Firstly, it is quite simple to implement a type-checker that can check types for exact correspondence, or for subtype compatibility with a given type. The type of the source object can be compared with that of the target variable to see if the former can be converted up to the latter, using subtyping rules like those we discussed in [1]. Secondly, code that has been checked once need never be checked again, or recompiled in new contexts. This is because the type system can never reveal more specific information about an object that is passed into a more general variable (which we have called the “type loss problem”), so the code need only be checked once over the most general type that it can accept.

Cite as:

Anthony J.H. Simons, “The Theory of Classification Part 20: Modular Checking of Classtypes”, Journal of Object Technology, Volume 4, no. 7 (September 2005), pp. 7-18, doi:10.5381/jot.2005.4.7.c1.