Microsoft .NET libraries write directly to fields instead of using properties?

Unless they're using automatic properties (where the backing field is hidden), I'd guess that most developers (including .NET Framework developers) creating a class will prefer to access private backing fields directly rather than going through the public properties.

Properties are really just a pair of methods. For example when compiling this class:

class MyClass

{

   private int _myInt;

   public int MyInt

   {

      get {return _myInt;}

      set {_myInt = value;}

   }

   // rest of the code

}

the property will get compiled (in MSIL) to:

public int get_MyInt()

{

   return _myInt;

}

public void set_MyInt(int value)

{

   _myInt = value;

}

Now, if this code is turned into native code for the underlying platform, there is a cost in accessing _myInt via these methods rather than directly which is why, I believe, most developers instinctively prefer the latter.

Having said that many experts believe that the 'just in time' compiler is clever enough to 'optimize away' the method calls and access the underlying fields directly and so there is no actual cost in using the properties. The trouble, of course, is that we don't know this for sure :)