Fast enough to drag

When you drag the time slider or a target solution, the numbers must keep up with your hand. We measured whether they will.

The Tote shows range and bearing to other platforms at the current time, and it must update within a tenth of a second of the slider moving. In Debrief v4 those numbers come from the Python service, one request per movement.

sequenceDiagram
    participant A as Analyst
    participant V as Debrief window
    participant P as Python service
    A->>V: drags time slider
    V->>P: Tote values at new time?
    P-->>V: ranges and bearings (about 2 ms)
    V-->>A: Tote updated
    Note over V,P: next request only once the last returns,<br/>always for the latest time

Each request took about 2 milliseconds, and never more than 7, even with twenty platforms: comfortably inside the tenth of a second.

Dragging a target motion analysis solution is harder, because every bearing in the period must be recomputed. With thousands of bearings, the straightforward code took a fifth of a second, too slow. Computing them all at once instead brought six hours of bearings at one a second down to under 40 milliseconds.

We measured on a development machine, so we will check again on an analyst's laptop before the target motion analysis work begins.