Open TorkelE opened 5 months ago
If we used the shift-index notation to describe the affect, we could actually handle cases like
discrete_events = 5.0 => [X + 1.0 ~ X + 10.0]
since it would be well-defined which X
to solve for, and easy to solve
discrete_events = 5.0 => [X(k) + 1.0 ~ X(k-1) + 10.0]
for X(k)
.
That's a pretty interesting proposal. We'd have to make validate that only a single shift is given but that would make that implicitness become explicit.
I have gone through the events in MTK. There are several cases where MTK allow your to build a system with a misformated event, and an error is not actually thrown until you create a problem or attempts to simulate it. Especially since it is a fit trick how to format the events, the errors should be thrown at the first possible time (when a system is created).