Question

I have started using C# 4.0 and loving the dynamic keyword. However, I am not sure what I am doing can be considered as good practice. Please see the code below:

static void Main()
{
    NoobSauceObject noob = new NoobsauceObject();

    dynamic theReturnType = noob.do(param);

    if (theReturnType.GetType().ToString().Contains("TypeOne"))
        theReturnType.ExecuteMethodOfTypeOne();
    else if (theReturnType.GetType().ToString().Contains("TypeTwo"))
        theReturnType.ExecuteMethodOfTypeTwo();
    else
        throw new ArgumentException("");
}

Is there a better way of doing this? I found the above method quite easy and have been using it but not sure if it is something I should stick with in the long run.

EDIT: If I were to do the same using .NET 3.5 or lower, or without the dynamic keyword, what would be a good implementation?

Thanks in advance!! 🙂

Answer 1

It looks to me like you’re just doing a type-test between two unrelated types. If possible, I’d look at polymorphism here, or at least: implementing a common interface. However, the following would be fine too:

var typeOne = theReturnType as TypeOne;
if(typeOne != null) typeOne.ExecuteMethodOfTypeOne();
else {
    var typeTwo = theReturnType as TypeTwo;
    if(typeTwo != null) typeTwo.ExecuteMethodOfTypeTwo();
    else throw new ArgumentException("somethingMeaningful");
}

However, my preferred option is:

var typed = theReturnType as ISomeInterface;
if(typed != null) typed.SomeMethod();
else throw new ArgumentException("somethingMeaningful");

where TypeOne and TypeTwo might use explicit interface implementation to expose the method on their API:

public class TypeOne : ISomeInterface {
    void ISomeInterface.SomeMethod() { ExecuteMethodOfTypeOne(); }
    public void ExecuteMethodOfTypeOne() {
        // ...
    }
}

(and likewise TypeTwo)

I can see no real use for dynamic here; For the return type of noob.do(param), object would be fine in the first example – or ISomeInterface would be even better.