Yes, that's what START did.
I wrote a single-pass S/360 assembler for teaching purposes.
One of the purposes was to assemble to location 0,
using absolute addressing. (Yes, I know that the low assresses on the real
machine were not available fot that purpose, but with START 0
& base register 0, instructions espressed in hex were simple,
consisting of op-code, operand register and an absolute address.)
From there it was a short step to instruction mnemonics and
base register of zero, and finally, with a base register no. >= 1
I/O was achieved using two spare op codes.
That was eminently faster than using the system's 360 assembler and linker.
----- Original Message -----
From: "Charles Mills" <[email protected]>
Sent: Wednesday, June 03, 2020 11:26 PM
I remember START and it definitely worked once upon a time. Here's what it was
good for:
In DOS/360 there was only a single address space containing the supervisor (nucleus) and three
partitions (regions). Only the first or lowest partition, called BG, was "full function" and that is
where most programs were run. It began at a fixed address following the supervisor, something in the
6K to 12K or so range. So most or all batch programs always loaded at the same fixed address. (Fixed
for the duration of a given configuration in a given shop.)
Let's say at your shop that address was x'2800'. If you were writing a main program and began it
with START X'2800' then your assembly listing exactly corresponded with the addresses in a dump. If
your program blew up at address x'3210' then that was offset x'3210' in your assembly listing.
Charles
---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
