On 3/22/19 10:28 AM, Glen Slick via cctalk wrote:
On Fri, Mar 22, 2019 at 9:59 AM Chuck Guzis via
cctalk
<cctalk at classiccmp.org> wrote:
At the expense of being boo-ed for this, could the original Rockwell
stuff perhaps have been assembled using a mainframe/mini-hosted
cross-assembler?
I'm aware of several situations where this was the case.
The date in the AIM-65 Monitor Program Listing header block in the
source code is Aug 22, 1978. That is less than 1 year after the
introduction date of the VAX-11/780. I suppose it still could have
been something that ran on a VAX by then, or a PDP-11 (or PDP-10?), or
some other mainframe/mini host if it wasn't self hosted on a Rockwell
6502 development system.
It's really just more of a curiosity issue at this point if anyone
finds a definitive answer.
Many cross-assemblers for early MPUs were written in (shudder!) FORTRAN.
There were several good reasons for this.
The first is that if you had a mini or mainframe, you were pretty much
guaranteed to have FORTRAN, which had been implemented under various
standards since 1966.
The other is that in the 70s, there was still a population of six-bit
character machines not using ASCII, not to forget the ones using EBCDIC.
So hard-coding character sets into programs that were supposed to be
portable over a wide range of machines was an issue.
I think some of the old FORTRAN code for PALASM may still be around, as
an example.
In that vein:
When I was tasked (1980) with producing a cross-assembler and cross-compiler
for the 68000 for our R&D sys, (Verex OS / Z language), the first operating target
was Motorola's 68000 emulator running on the campus mainframe (MTS on Amdahl / 370).
(Followed by hardware, which was a 68000 exerciser board or a bare SUN-1 processor
board).
I'm pretty sure there was also a 68000 cross-assembler from Moto on the Amdahl,
although I'm not sure whether I used it or not, might have to confirm it's output
with the output from mine.
IIRC the Moto programs were written in Fortran (oops, FORTRAN).