|
|
On 05/23/2006 10:55:45 PM, Javier Burroni wrote:
> (previously sent to quantlib-dev)
I know---answer are not always quick :)
> I'm reading the TimeGrid code and I've noticed that the constructor
> of this object using two iterator is not a particular case of the
> constructor using two iterator plus a step parameter (set to zero),
> and the object behaves different. Is there any reason to this?
Historical reasons, I think. There might have been a use for the
behavior when steps = 0, but it escapes me now. In his youth, a
developer I won't mention used to try and foresee all possible ways in
which users might want to use a class; he also had a habit of encoding
such behaviors in magic numbers. He is older and wiser now :)
> Also, in the constructor wich use the end and a Step, the value 0.0
> is not set as the first mandatory time.
As above. Had I the time, I would go and look at where mandatory times
are used to see whether any code expected 0 _not_ to be there. Anybody
on the list is willing to make a code review?
> ps: the use of this template oriented c++ you are using is awesome :)
Thanks.
Later,
Luigi
----------------------------------------
Everything can be filed under "miscellaneous".
-- unknown
|
Locked
