Are DataTypes needed?

[mtf90]

Do we actually need DataType as a standalone entity, or could the model be simplified by using the respective Class<?> as type constraint directly? The only place where I could see it being useful is in the teachers definitions (Map<DataType, Theory>). Having it as an explicit entity would allow one to use different theories for same-typed data types simultaneously (e.g., an integer EqualityTheory for one parameter type and an integer InequalityTheory for another parameter type). Is this something that should be supported?

[FredrikTaquist]

The DataType entity allows us to distinguish between types that happen to have the same Class<?>, but should not be treated as the same data type. For example, say that you have a userID and PIN data type, which can both be represented as integers, but are clearly not the same type.

Also, as you suggested the question, having a separate DataType allows us to have different theories for the different data types, even if the underlying Class<?> is the same. Since we plan to support more complex theories, such as learning with successors (guards of the form s2 == s1 + 1) it is useful to have this functionality.

[mtf90]

Ok, then it makes sense to leave the concept as-is.