Somewhere in my travels on the web (and I think it was here on Wiki) I saw a reference to a pattern called something like InitializeNotAssign.

It concerned the use of the C++ constructor member initialisation list. e.g. given a class
  class Myclass
  {
  public:
	int a,b;
	Myclass();
  };
This
  Myclass::Myclass() : a(1), b(2) {}
is preferable to 
  Myclass::Myclass() { a = 1; b = 2; }	
Someone has asked why and I wanted to throw the pattern into the discussion. Anyone seen it?

-----------

OK, I found it again. It's specific to C++ so I won't quote it here. It was in the May 96 C++ Report, and article by JimCoplien called "After all we can't ignore efficiency".

For those interested the article is available at
http://www.sigs.com/publications/docs/cppr/9605/cppr9605.c.noname.html

--KeithDerrick

----
CategoryCpp