[Tutor] Problem with default arguments for function

[Dick Moores]

At 05:10 PM 11/11/2007, Michael H. Goldwasser wrote:


>Dick,
>
>Another typical strategy is to use some prescribed special value for
>the precision parameter to designate the desire for full precision.
>For example, since precisions should presumably be positive, one could
>design this function as:
>
>def fact(n, precision=15):
>     """compute n!.
>
>     precision    the minimum desired precision.
>                  If -1 is specified, computed to full precision.
>     """
>     # ...
>     if precision == -1:
>         precision = n * 10  # insures that for n < 1 billion, ...
>     # ...
>
>
>If you are not happy with the oddity of -1 (or in cases where -1 might
>be a legitimate parameter value), you can pick the flag from a
>different data type.  In this case, perhaps None would be a more
>natural way to say that you do not want any limit on the precision.
>So this could be coded as
>
>def fact(n, precision=15):
>     """compute n!.
>
>     precision    the minimum desired precision.
>                  If None is specified, computed to full precision.
>     """
>     # ...
>     if precision is None:
>         precision = n * 10  # insures that for n < 1 billion, ...
>     # ...
>
>Looking at your examples, this should (unteste) behave as:
>
># 1 (precision default overridden and set to 20)
>  >>> print fact(50, 20)
>3.0414093201713378044e+64
>
># 2 (without explicit value, precision defaults to 15)
>  >>> print fact(50)
>3.04140932017134e+64
>
># 3 (explicitly says not to limit precision)
>  >>> print fact(50, None)
>30414093201713378043612608166064768844377641568960512000000000000

Beautiful! Thanks!

Dick