Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 12 Feb 2020 01:24:38 -0600
From:      Scott Bennett <bennett@sdf.org>
To:        freebsd-questions@freebsd.org
Cc:        "Steve O'Hara-Smith" <steve@sohara.org>
Subject:   terminology and history (was Re: Re  updating BIOS)
Message-ID:  <202002120724.01C7OcSW005991@sdf.org>

Next in thread | Raw E-Mail | Index | Archive | Help
     On Sun, 9 Feb 2020 08:41:11 +0000 Steve O'Hara-Smith <steve@sohara.org>
wrote:

>On Sun, 09 Feb 2020 02:09:59 -0600
>Scott Bennett <bennett@sdf.org> wrote:
>
>>      The first part of the above, mispunctuated pair of sentences is
>> correct, but the latter part is not.  FreeDOS, like PC-DOS and MSDOS
>> before it, is/was not an operating system, but rather a more primitive
>> creature known as a monitor system.
>
>	The DOS part of those names is an abbreviation of 'Disc Operating
>System' - clearly at the time they were considered operating systems even

     They may have been considered that by amateurs from the ham radio
community or by the ignorant twerps like Bill Gates et al., who closed
their eyes, ears, and minds to all that was already well known to people who
had been working in the field for years.  That Micro$lop has always misnamed
things from time to time should come as no surprise to anyone here.  Calling
a radio a television does not make it a television.

>though they started life as near clones of CP/M (Control Program/Monitor).

     And that was more correctly named, though/because it was *not* named by
Micro$lop.

>IBM 360 mainframes didn't have virtual memory, processes or any of the

     Virtual memory support was a very late addition to operating systems.
Before that time, all operating systems and monitor systems were real memory
systems only.
     While most System/360 models had no hardware features to enable operating
system support for virtual memory, the model 67 had an extra foot or two of
cabinetry filled with the Dynamic Address Translation unit and the CPU could
run in standard PSW mode or Extended PSW mode.  When in EPSW mode, address
translation might be enabled or not.  DOS/360 and OS/360 were real memory-only
systems, but TSS/360 had full virtual memory support with most of the features
that later virtual memory systems had.  Although offered by Cambridge University,
rather than IBM, CP-67/CMS provided virtual machine support.  Some installations
ran OS/MFT, OS/MVT, or later OS/MFT II in one or more VMs for batch processing
and, for interactive users, CMS in users' VMs.  All of this was around by the
late 1960s.  In the mid 1970s, IBM adapted CP-67/CMS for the enhanced System/370
models as VM 370/CMS and began releasing VM versions of its other systems (DOS/VS,
OS/VS1 (i.e., virtualized MFT-II), and OS/VS2 SVS, and OS/VS2 MVS).

>protections you mentioned, it didn't even have anything that would be

     "Processes" were called "tasks" in IBM's operating systems.  MVT allowed
subtasking, the closest analogue of which in UNIX would probably be threads.
     All System/360 models (except the 20 and perhaps the 25) either had
storage protection standard or could be ordered with it as an option, and
were supported by DOS/360 and OS/360 if present.  I don't know whether TOS/360
provided storage protection support,  TSS/360 required storage protection, but
that was standard on the 67 (I think on all models 50 and higher, actually).  It
also used storage protection in a different manner from the other systems because
it could effectively handle the original protection functions much better through
memory mapping, while the storage keys could help in managing other matters.

>recognised as a filesystem today (it had record oriented datasets) - but

     They used the standard Volume Table of Contents, which was both volume label
and file system.  (A file system need not provide a nested directory structure.
IIRC, VMS also had a single-level file system.)  They provided access methods,
OS/360 having several more than DOS/360.  OS/360 also had a systemwide catalogue
of data sets for those that the user(s) told the OS to keep track of in a central
location.  They provided tape labelling and recognition, which UNIX systems still
do not provide..
     Although OS/360 could be installed with one of three options (PCP, MFT,
or MVT), which determined to what degree the system would be a multiprogramming
system, DOS/360 was a partitioned, multiprogramming system, roughly analogous
to OS/360 MFT.  (MFT only had one job scheduler running and had other scheduling
limitations, too.)  In later years, MFT was replaced by MFT II, which allowed
many more partitions and with job schedulers running in partitions that were
large enough.

>OS360 was definitely considered an operating system.
>
>	[MS/PC/DR/Free]DOS was a lot more like a mainframe batch operating

     No, that was my point.  They were all like monitor systems (e.g., IBM
1620/1710 Monitor I).  They did almost nothing for the user or program except
for loading an executable program from a disk drive and accepting a return of
control when the application program ended, so that the next program could be
loaded and control transferred to it, just like {MS,PC,DR,Free}DOS.  That
was a big advantage over having to load a standalone loader on a deck of cards
preceding every object program deck one wanted to run, but it certainly was
inferior to an operating system.

>system than a multi-user multi-tasking operating system such as Multics or

     Some operating systems embodied no concept of distinct users.  For example,
if one had not set up the use of accounting features in DOS/360 or OS/360, all
users were effectively one.  With accounting in use the resource usages could be
kept separate for, say, billing purposes, but that was basically all.  As far as
privileges were concerned, a program ran in user (non-privileged) or supervisor
(privileged) mode, but again there was no distinction among users.  Very little
outside of the supervisor itself ever was allowed to run in supervisor mode.
(One exception was spooling systems like HASP, ASP, POP, GRASP, etc., which often
came with their own SVC routines that made it possible for them to intercept unit-
record I/O to be spooled.)

>unix, but hijacking the term operating system to mean only the latterm, and

     I've been using the terms consistently since 1966.  Said hijacking of terms
was what I was complaining about.
     UNIX was very much a latecomer.  There were many monitor and operating
systems before its first appearance on the scene.  Every manufacturer offered
its own proprietary system(s).  UNIX also was a real memory operating system for
many years until after the VAX-11 models were released a decade or more after
the 360/67 and at least three years after the 370s appeared that had hardware to
enable virtual memory features of operating systems.  (It is worth recalling here
that VAX-11 stood for Virtual Address Extensions [to the PDP]-11.)


>that only with hardware supported isolation mechanisms is revisionist. I
>recall working on a unix(ish) system in the late 1980s that didn't have
>hardware memory mapping or protection, or even fsck which made recovering
>from (the frequent) crashes rather tedious (icheck, ncheck ...).

     ...which some filesystems did not/do not need. :-)
>
     You have it backwards.  Erasing prior computing history would appear to be
"revisionist".  You're commenting about a time two decades after the appearance
of virtual memory features becoming available in both hardware and software, so I
don't know what your point here is supposed to be.  Furthermore, one of the
TSS/360 developers once described to me how they were testing that system in a
software simulator of the model 67 on a model 40 a year or two before the 67
hardware was available and working correctly.  Note, too, that the VAXes, when
they finally appeared, had a design flaw in that they had no reference bits to
tell when pages had been accessed for reads, so the CSRG had to simulate
reference bits in the kernel.  That was kind of inefficient in comparison to
reference bits set by the hardware, but far better than not keeping track at all.
Why DEC made such a mistake after better examples had been on the market for a
good decade I do not know.


                                  Scott Bennett, Comm. ASMELG, CFIAG
**********************************************************************
* Internet:   bennett at sdf.org   *xor*   bennett at freeshell.org  *
*--------------------------------------------------------------------*
* "A well regulated and disciplined militia, is at all times a good  *
* objection to the introduction of that bane of all free governments *
* -- a standing army."                                               *
*    -- Gov. John Hancock, New York Journal, 28 January 1790         *
**********************************************************************



Want to link to this message? Use this URL: <http://docs.FreeBSD.org/cgi/mid.cgi?202002120724.01C7OcSW005991>