Jean-David Beyer wrote:
Another way of looking at your statement would be if you thought I meant a
language in which to write operating systems. I think it would be possible
to do it in BASIC, but I sure would not want to try it. I know ALGOL 60
compilers were written in ALGOL 60, even though it was by no means ideal to
do so.

Not always. The Burroughs machines of that era (late 1960's) were true
stack processors and Algol suited them perferctly. The same with the HP
3000 computers of the early 1970's which also used some variant of Algol
as their "machine language".


Linux was first written in assembler for the PDP7
(and later the PDP11), and was converted to C in the early to mid 1970s.


UNIX not LINUX!!! Linux did not exist in any form until the early 1990's
and it started life as a terminal emulator.



Later Steve Johnson wrote the portable C compiler that made it practical to
port the UNIX OS from one kind of machine to another. I believe they ran it
on Interdata machines, and they for sure ran it on System/360 machines and
Amdahl machines. Of course now it will run on any machine that have the
right resources that people care enough to port it.

The first 360 operating system was TOS (tape operating system) and it
was written in BAL (assembly language), the same with OS/360 (PCP, MFT and
MVT), DOS/360, CP/67 and its CMS and ACP (Airline Control Program).

The 370 variants (DOS, DOS/VS, DOS/VSE from DOS), VM/370 and later VM/XA
(from CP/67), SVS (from OS/360 MFT) and MVS (from OS/360 MVT) and the 370
version of ACP were still written in BAL. I don't know about the others, but
MVS was 80% MVT code.

In 1980 so I was pretty familar with both MVS and VM source code,
having worked on additions to MVS's open/close/end of volume processing
and lots of work on VM (which was open source) from 1978 until the late
1980's.

MVS was "open source" in that you could with some legal agreement get a
copy of the assmebly listings on microfiche and JES (the part that handled
the printers/etc) was released in source code.

VM (and CMS) was released with source code and all fixes were provided as
source code. You had to apply the patches, reassemble and link the modules.
At some time (around 1987) IBM announced a vector processor for the 308x
series of computers and the modules that supported it were object code
only (closed source). It created quite a furor, although 99% of those
computers were not sold with one.

I don't think that UNIX (IBM called it AIX) was available for either the
360 or the 370 in that era, from what I remember it came after AIX for the
RISC 6000 which would place it in the 1990's.

Amdahl machines ran either MVS or VM/370. By that time SVS was gone and no
one ran DOS on "big iron" unless it was in a virtual machine under VM.

BTW, the predecessor to the HP 3000, the 2100 series had a DOS (disk
operating system) and a standalone BASIC timesharing system. Source code
was available for them. I used to have a binder with the source code
listings for the BASIC system and remember making patches and reassmebling
the DOS. This was the late 1960's-early 1970's.

I also remember wanting to try APL in the late 1960's and having only
access to a standard teletype and the aformentioned HP BASIC system,
writing an APL interpreter in basic. It was fine except that you had to
spell out each of those funny operators.

Geoff.

--
Geoffrey S. Mendelson, Jerusalem, Israel N3OWJ/4X1GM