I wrote this source listing using the Bibo-Assembler. 
You may ask why - since I normally prefer the Mac/65. Well, 
I admit there are several things which Mac/65 cannot do. 
Among these are the commands RUN (label) and LIST (label). 
In it's beginning the HDI consisted of two test boards. 
On one there was the SIO chip and on the other there was the floppycontroller (FDC). 
Those boards were made to fit into the cartridge port. This way I was able to directly 
address those chips from assembly language and hence test the routines easily.


Maybe someone noticed that the HDI I/O ports all are in the $4xxx range? No? 
But this is intentionally, too. When the final HDI board was done there 
should be a way of easily testing the HDI operating system without steadily 
burning EPROMS and working almost blind. Doing so the HDI CPU socket was 
connect (almost) 1:1 to the XE CPU socket, ROM, RAM and CPU removed from the HDI. 
A little circuit provided switching off all chip select signals in the XE within $4000 
to $7FFF via CASINH. Thus test routines could be run directly from the Atari enabling 
hardware register access from a machine language monitor and it provided debugging 
options to the HDI (I was able to run the HDI OS in the XE memory).