Fair enough, as a computer scientist I got tought to use the Neumann definition, which includes zero, unless stated differently by the author. But for general mathematics, I guess it's used both ways.