> On Aug 1, 2017, at 8:42 PM, Emil Constantinescu <[email protected]> wrote:
> 
> On 8/1/17 7:29 PM, Barry Smith wrote:
>>> Well, there are situations in which you want to do a certain number of 
>>> fixed steps.
>>    Really, I think users ALWAYS want to get to a particular time; it is only 
>> due to flaws in the software that they may think they need to (or due need 
>> to) control things with a fixed number of steps.
> 
> Yes, but you cannot always guarantee that. E.g., test convergence of whatever 
> problem when using euler, beuler, rk 4 (very popular methods). You cannot 
> force them to stop at a certain time unless you fix the number of steps 
> because none have adaptors.

  Exactly my point. They WANT to use a particular time but the software isn't 
good enough to allow them to.

  How about the AI based software take the time they request T divide by dt and 
then set that as the number of time steps to use. This gives a uniform 
interface for the user and still gives the ending time what you want. My point 
is having the USER set the number of time steps is a bad API.

   Barry

> 
> Emil

Reply via email to