Why is VEserial.begin() called within the read() function?

[Hyperdimensionals]

Hello,

I'm curious why the VESerial.begin() function is called within the VEdirect.read() function. My understanding is that serial.begin() is typically called once when initially establishing the serial connection. Does this make it more resilient if the serial port gets temporarily disconnected?

I also noticed that in the VEDirect::begin() function, VESerial.begin() is called with the Baud rate hard-coded (VESerial.begin(19200), while in the VEDirect::read() function, it's called with the macro constant (VESerial.begin(VED_BAUD_RATE)).

[Hyperdimensionals]

I tested VEDirect::read() without VESerial.begin() and it worked fine. As I understand it, once serial ports are open on Arduinos they always 'exist' to transmit and receive regardless of whether there is a physical connection on the other end. I suppose the advantage could be not having to call VEDirect::begin() initially?

Also, looking at the VEDirect::begin() function:

uint8_t VEDirect::begin() {
	// Check connection the serial port
	VESerial.begin(19200);
	if (VESerial) {
		delay(500);
		if(VESerial.available()) {
			VESerial.flush();
			VESerial.end();
			return 1;
		}
	}
	return 0;
}

Doesn't flush() only apply if this library were transmitting data? I also don't understand the reason for including end().