"Simple establish a timer interrupt of e.g. 1ms and count down so many bytes as simultaneous delays you need. "

Yes I also do that quite often. But the downside is, you still have to CHECK or POLL to see if the counter(s) has reached zero. This is additional software overhead. Whereas with the built in hardware timer or the PCA, the MCU can let it know with an interrupt.

Are we in agreement or I am missing something ?

Raghu