Question about AP_PID example


I am looking at the AP_PID source code and its example.

This is the example code:

[code]void loop()
// setup (unfortunately must be done here as we cannot create a global AC_PID object)
uint16_t radio_in;
uint16_t radio_trim;
int error;
int control;
float dt = 1000/50;

// display PID gains
hal.console->printf("P %f  I %f  D %f  imax %f\n", pid.kP(), pid.kI(), pid.kD(), pid.imax());

// capture radio trim
radio_trim = hal.rcin->read(0);

while( true ) {
    radio_in = hal.rcin->read(0);
    error = radio_in - radio_trim;
    control = pid.get_pid(error, dt);

    // display pid results
    hal.console->printf("radio: %d\t err: %d\t pid:%d\n", radio_in, error, control);


I do not understand why dt is set as 1000/50. As I understand, dt should be the sampling period i.e. from the beginning of the loop towards the end of the loop, which is approximately 50ms+ in this case. Can anybody explain this to me?

I think the 1000/50 is an error. Looks to me like the AC_PID library takes its dt in seconds. The main loop runs every 50 milliseconds, so the correct dt would be 50/1000 = 0.05 sec.