It's a pretty common occurrence to see a programmer completely disregard regular expressions, instead implementing his own validating code. A handful of the Do-It-Your-Selfers will even implement their own substring searching algorithm with all sorts of varieties of loops and the like. David Conrad's colleague managed to not only skip the built-in parseFloat method, but use regular expressions, an array, and a loop to do so ...

function testNum(strNum)
{
  if (strNum.length == 0) return false;

  var theChar, idx, isValid;
  var theRegExp = /[0;1;2;3;4;5;6;7;8;9;.;-]/
  var charArray = strNum.split("")
  for (idx = 0; idx < charArray.length; idx++)
  {
    theChar = charArray[idx]    
    isValid = theChar.search(theRegExp)   
    if (isValid == -1) // a character was found that was not numeric
    {
      return false;
    }
  } 
  return true;

}