This question is subjective but I was just curious how most programmers approach this. The sample below is in pseudo-C# but this should apply to Java, C++, and other OOP languages as well.

Anyway, when writing helper methods in my classes, I tend to declare them as static and just pass the fields if the helper method needs them. For example, given the code below, I prefer to use Method Call #2.

class Foo
{
  Bar _bar;

  public void DoSomethingWithBar()
  {
    // Method Call #1.
    DoSomethingWithBarImpl();

    // Method Call #2.
    DoSomethingWithBarImpl(_bar);
  }

  private void DoSomethingWithBarImpl()
  {
    _bar.DoSomething();
  }

  private static void DoSomethingWithBarImpl(Bar bar)
  {
    bar.DoSomething();
  }
}

My reason for doing this is that it makes it clear (to my eyes at least) that the helper method has a possible side-effect on other objects - even without reading its implementation. I find that I can quickly grok methods that use this practice and thus help me in debugging things.

Which do you prefer to do in your own code and what are your reasons for doing so?