make sure you are in 8 bit mode (read the manual about writing the mode byte twice since you wake up in 4 bit mode).

Ignore the "busy". Interface with a "write only" xface and write to the LCD with timer interrupts or with a routine that guarantees that you will not provide data with an interval less than the max command latency which is (if i recall correctly) is 100us. for "clear and locate cursor home". I have done this for every LCD i have ever used. This technique allows simple printer port interfaces or use of a simple latch which free's up the port you may be using for alternate functionality.

regards,
p