What is the difference between variable = 0 and variable = nothing in .NET?

Question

What is the difference between:

Dim intVal as integer
intVal = 0

and

intVal = nothing

I have read from MSDN that nothing sets a value to 0 as well:

From MSDN:

Public Structure testStruct
    Public name As String
    Public number As Short
End Structure
Dim ts As testStruct, i As Integer, b As Boolean
ts = Nothing 
' The preceding statement sets ts.name to "" and ts.number to 0.
i = Nothing 
b = Nothing 
' The preceding statements set i to 0 and b to False. 

Answer 1

Nothing is usually used for object references, if you assign it to a value type it initialize it with the default value. In .NET world, default values are simply a block of memory full of zeroes. So, an integer assigned to nothing will be 0, a struct assigned to nothing will be a structure filled of zeroes.

A null reference is a special reference, it means, no reference, and is expressed, indeed, as a pointer with all zeroes.

In C# you can use the default(TYPE) to obtain the same result, it is used to assign zero to whatever you want.

String are references type, like object, so strings will not be assigned to "" but to Nothing, strings are not value types.

Dim s as String
s = Nothing

Console.WriteLine Object.ReferenceEquals(s, Nothing) ' Will print true
Console.WriteLine Object.ReferenceEquals(s, "") ' Will print false

Strings are still reference types, so when you assign it to Nothing you will set the reference to zero.

Answer 2

In .Net an Integer is a value type and therefore cannot be null.

If you create a variable of type integer and don't assign it, it will be 0 because of the way the language works.

It's not hyperbole to say that some of the most irritating/prevalent bugs in modern computing have been caused by un-initialised variables so the .Net team made the decision to ensure that all value types don't suffer from that problem.

It is bad practise though to assume that the compiler/JIT will resolve these variables to a known value such as 0 for integers or False for booleans, so you should always use your first example.

But to answer your question clearly, there is no difference, but it's best to use the first example.