Revision history [back]

In case you retrieve the state variable from the API, why don't you calculate it in an 'a posteriori' pattern, after the value of the state variable is available? You could calculate it as a divided difference, since the time step is known.