A number of issues are explored concerning the notion that a data type is a set of values together with a set of primitive operations on those values. Among these are the need for a notation for iterating over the elements of any finite set (instead of the more narrow for i := 1 to n notation), the use of the domain of an array as a data type, the need for a simple notation for allowing types of parameters to be themselves parameters (but in a restrictive fashion), and resulting problems with conversion of values from one type to another.
Read full abstract