[Milsurplus] IBM PC

[Dave Emery]

On Thu, Sep 16, 2004 at 05:43:36PM -0400, J. Forster wrote:
> Ray Fantini wrote:
> 
> . The IBM had DOS, ok so it's a Microsoft reverse engineered version of CPM, but
> once again everyone else used it, well except for Mac people.

	As an old software type I think I can speak on this ...
Microsoft bought MSDOS from a small company in Seattle that had
developed it as a CPM clone/competitor.    Legend has it that IBM
managers in semi-covert mode (they did not want the world to know about
their PC project) visited the company that actually developed early
MSDOS well before Microsoft had anything to do with it  and the owner
blew them off so they asked Bill Gates who they were also visiting to
buy a Basic for their machine if he could supply a DOS too and he
immediately bought the rights to MSDOS from the guys across town and
struck the infamous licensing deal that gave him the right to sell it
independently of IBM that made his vast fortune.

> DOS was around YEARS before. There was certainly a stripped down version of DOS,
> loaded from paper tape onto a head per track hard disk with a capacity of 64K 16
> bit words and up in about 1969. It was the DG 4019A and used an Alpha Data
> drive. They were available in sizes up to at least 512 KW. Small, but fast. No
> head seek time, just the latency. That DOS would run in something like 8 KW.

	John is confused here.   The term DOS stood for Disk Operating
System and was used by quite a few different and unrelated companies to
name their operating system for use with disk hardware back in the 60s
and early 70s.   I know of unrelated operating systems called DOS for DG
hardware, DEC hardware, Prime hardware, Wang hardware, early Intel
hardware, and several others.  The DG DOS, RDOS, and RTOS  has nothing
whatsoever to do with MSDOS (later shortened in most people's minds to
DOS).  There is no historical relationship - MSDOS is a clone of CPM, DG
operating systems were quite different and unrelated.

	Very early minicomputers and even some small mainframes did not
have disks (either hard drives or floppies) but ran off of mag tape
(remember DECTAPE ?) or in the case of minicomputers programs loaded
from paper tape.

	This necessarily imposed a lot of design constraints on operating
systems (sometimes called monitors or loaders back then) which had to
operate out of very limited memory with no random access devices for
storing files on.   When disks came along they made random access
storage of files practical, which allowed all kinds of new and wonderful
things such as directories ("folders"), directory trees, easy creation,
addition to, shrinking of and deletion of files, and programs that
opened multiple files at once, and soon complex databases of all kinds
and sorts including modern relational types.

	And disks also made it possible to dynamically load and unload
programs and parts of programs as they were needed, which made it
possible to create both operating system components (such as command
line interpreters) and large user programs (such as compilers) that were
much much larger than the very limited memory of early machines (though
tiny by comparison with their modern descendants which are often 1000
times bigger) and still run them efficiently and responsively.

	Adding disks created a revolution in operating system design and
for early computer companies this meant developing and offering the
shiny new Disk Operating System for their hardware.    And very few
early disk operating systems would run usefully without disks though a
few tried.

	All of this experience with mainframe and minicomputers repeated
with microcomputers based on microprocessors in the early to mid 70s,
and users and developers very rapidly discovered that a computer with
some form of "high" capacity random access storage was vastly more
useful than one that had only serial access storage such as cassette
tape or paper tape.

	In the case of early pre-PC desktop computers the device the
market settled on was the floppy... (not around in the early days of
minicomputers or mainframes).   It rapidly vanquished the tape cassette.

	But as an interesting footnote to history - IBM was unsure what
users would want in their PC, so the first machines did have a cassette
interface and only an optional floppy...


> 
> There were two versions, RTOS (Real Time Operating System) and RDOS (Real Time
> Disk Operating System). I first came across RDOS with the Diablo 31s which were
> pizza oven like 2,5 M hard drives (IBM 2315s).

	I worked on RDOS and RTOS while at DG in the 70s.   It was a lot
more sophisticated real time operating system than MSDOS ever was
(particularly early versions of it).

	Its immediate ancestor at DG was called DOS (and developed more
or less single handedly in 1968-70 by my immediate boss at DG in the mid
70s).


> IMO, S-100 and CPM were significantly later.

	But disk operating systems went back much further than DG DOS or
RDOS or DEC OS-8.   Early work on mainframes in the late 50s and 60s and
such pioneering machines as the Whirlwind of SAGE/MIT fame included
almost all the features and capabilities of even DOS 2.0 or 3.0 and some
that MSDOS never incorporated.



-- 
   Dave Emery N1PRE,  die at dieconsulting.com  DIE Consulting, Weston, Mass 02493