Why does interpolation cycle? about mainloop.js

[IceCreamYou]

That value is frameDelta / simulationTimestep. frameDelta is the elapsed time that hasn't been simulated by update yet, and simulationTimestep is the amount of time processed by each update. So essentially the value is the percent of an update that hasn't been simulated yet. If the amount of time between frames is anything other than exactly simulationTimestep, which will always be the case in practice for at least some frames, then the difference will cause a change in the interpolation amount. At ~60fps the change is very small in each frame, so you see the interpolation amount creeping up slowly from 0 to 1. Once it reaches 1, there is enough accumulated unsimulated time for another update, bringing the value back down to zero. However, the cyclic pattern is not guaranteed or monotonic; it just happens to be the case that with your demo code and under the conditions in which you ran your demo, the time between frames happens to be consistently slightly more than simulationTimestep. Under other conditions the time between frames could vary, or it could be below simulationTimestep, or it could be different enough (e.g. 25ms) that a pattern is not easily discernible.

[dakom]

Followup question with the same demo - why does delta always increase in begin?

The docs define this as: "The total elapsed time that has not yet been simulated, in milliseconds"

I don't quite get what it means that there is a constantly growing value of "elapsed time that has not yet been simulated"

[IceCreamYou]

delta does not always increase in begin. Not sure what you're referring to. When I open your demo, delta hovers in the 16-33ms range, which is normal for a program running at ~60fps; it basically means there's one frame's worth of time to process just before each frame, which is what you'd expect.

[dakom]

Ah - I saw it climbing and didn't wait long enough to see it come back around ;)

thanks!