You would have to provide more details for me to answer, but read chapter 8 on user-defined types. That may help.

Thank you gentlemen for diving further and expanding on my question.

I was able to figure it out, the types had to be explicit as to what you pass into a collection of a custom type.

i.e. Haskell CAN'T infer [MyType 8, MyType "fib"] but CAN infer [IntType 8, StringType "fib"] even though the data MyType is of either type IntType or StringType. Seems as though Haskell always wants the most explicit type which is interesting because I would think during compile time Haskell can interpret that MyType is of two types and will proceed to check which of the two types an integer or a string will fall under when it's explicitly declared but it doesn't, I guess Haskell isn't truly lazy :p

The book doesn't go into this but I managed to figure it out.

bfazeli wrote:

Thank you gentlemen for diving further and expanding on my question.

I was able to figure it out, the types had to be explicit as to what you pass into a collection of a custom type.

i.e. Haskell CAN'T infer [MyType 8, MyType "fib"] but CAN infer [IntType 8, StringType "fib"] even though the data MyType is of either type IntType or StringType. Seems as though Haskell always wants the most explicit type which is interesting because I would think during compile time Haskell can interpret that MyType is of two types and will proceed to check which of the two types an integer or a string will fall under when it's explicitly declared but it doesn't, I guess Haskell isn't truly lazy :p

The book doesn't go into this but I managed to figure it out.

Good to know that you managed to figure it out.

Actually Haskell types and their values are quite different than any other languages. Thus, we can not expect IntType to be interpreted dynamically as a subtype of MyType. In fact, IntType is not a type at all. It is (along with the parameter) just a value out of all the possible values of MyType. Just like 2 or 3 or 4 are the values of a type int. Therefore, you have to be explicit about what value (i.e. IntType _ or StringType _) of a type (MyType in your case) you desire to pass. At the receiver function, we can consume this variation with the expectation that it would be of MyType type, and then handle them for each possible values (or functions as in your case) separately.

This is somewhat similar to the Interface in Java. That is, given a function compare(Comparable obj1, Comparable obj2) where Comparable is an interface, we cannot make a call as compare(new Comparable(), new Comparable()). However, what we can do is, pass on those objects that implement Comparable interface OR add the required implementation during the instantiation at the method call itself. In both the cases, what we are doing is passing on a value (out of all the possible values) of a type Comparable and not the type itself.

You are correct that book didn't cover this scenario, however there is one line where he clarifies the difference between type and the value (or value constructor):

We couldn't write a type declaration of Circle -> Float because Circle is not a type, Shape is.

That is how I understand types in Haskell, hope it will help.