True, using C I did suffer from that problem until I realized it, however with this language I have not had that problem, I suppose I would do it internally with Swordfish.
Regarding the "NOPs" when placing those fixed values on the slow displays, the times were reduced to a delay of 30uSec. for the DIO line and a fixed delay for all (slow and fast) of 1uSec. for the CLK.
Therefore I use a fixed delay2 of 1uSec for the CLK line for slow and fast displays.
For the DIO line I use a delay1 of 1uSec for fast displays and a Delay1 = 30uSec for slow displays.