Oleg T. wrote:
-------------------------------
Is that using some constant address values instead of values that could be
calculated by linker from function/structure identifiers?
If so, it's clear.

= = = = = = = = = = = = = = = =
That's what I meant. In other words, if you use constant values it is tedious to make changes. I had the man-machine interface running on a Ampro CPM system though that was masked from the user. I did all the assembly and linking on the same Z-80 system we shipped for the console. :)

The variable lengths were supported because the menu structure for drill-down information was not specified precisely. The order on subscreens and the count of items could change upon the customer's request so the structure made it easy.

I invited them to come down and view a simulation and suggest changes. I ran the Z-80 system with the dumb terminal in the conference room and by the time they decided on how they wanted it re-worded or re-ordered, I had moved the menu data structure around and recompiled so they could immediately see its operation. Its was a smooth meeting. :)

Actually, if you took someone's suggestion on how to organize your menu structure... you'd miss all the fun.

aka j

PS: This contract was an upgrade of a unit we had sold the Air Force a few years before. I went to the Air Force base to do a quick service job for them and heard everything they hated about the previous design. The engineer that did the original design left the company so I was able to make a lot of improvements without him fighting about it.

He had given them a man-machine interface where the operator had to enter commands like a basic microprocessor monitor program: A120,60,2 sort of format. It was obtuse.

The most amazing thing the previous engineer got away with was a 2Mhz Z-80 run about 40 feet to peripheral control cards through cables and repeater structure. He designed to 4Mhz but couldn't get it to work. The arrival of data, address and control signals was completely messed up. I fixed that and threw out his man-machine interface and processing algorithm unit and replaced it with the little Ampro. I used a SCSI interface to connect the two Z80's. They had to pipe a lot of information at fast rates.