From: cross@spitfire.i.gajendra.net
In article <878qpd130b.fsf@example.com>,
Salvador Mirzo wrote:
>cross@spitfire.i.gajendra.net (Dan Cross) writes:
>> In article <87tt83besr.fsf@example.com>,
>> Salvador Mirzo wrote:
>>>Grant Taylor writes:
>>>
>>>> On 7/18/24 10:07, Paul Edwards wrote:
>>>>> For 35+ years I have wondered why there was no MSDOS for the mainframe.
>>>>
>>>> The answer is in the name.
>>>>
>>>> MS-DOS
>>>>
>>>> Microsoft DOS
>>>>
>>>> Micro
>>>>
>>>> micro-computers are the smallest end of the system with mainframes and
>>>> supers at the other end of the system.
>>>>
>>>> IBM provided a Disk Operating System for early and / or smaller
>>>> mainframes.
>>>
>>>And why is /Disk/ Operating System? What's so /disky/ about it?
>>
>> Simple: it drove a system with a disk. Most early mainframes
>> didn't have disks, so once they came along, system software had
>> to evolve to meet the needs of new hardware.
>>
>> IBM's DOS/360 was pretty anemic compared to it's flagship OS360.
>> But it was built as something of a stopgap because OS was behind
>> schedule.
>
>Thanks! Changing the subject a bit to the history of DOS, if that's
>okay.
Yes, of course.
>I was not quite aware that there was a mainframe DOS in the IBM
>world. So it seems to me tbat Microsoft found the DOS made by ``Seattle
>Computer Products'' the right choice to buy because they wanted to
>produce a system for IBM micro-computers---it makes sense in sort of
>keeping the same user interface. But this strategy assumes that the
>users of micro-computers would be the more or less the same users as IBM
>mainframes. Am I imagining things correctly here and did the strategy
>really make sense? (It could also be the case that Microsoft just
>didn't have any other option.) (Background: I've watched the film
>``Pirates of Sillicon Valley'' a long time ago. That's how much I know
>about the history of MS-DOS.)
Well, I would urge some caution here; I don't think that DOS/360
had much resemblence, if any, to MS-DOS: it was a batch system
for very low-end mainframes in the IBM 360 line. The name clash
is just a coincidence. At the time, lots of manufacturers were
starting to introduce "DOS" systems, since disks were relatively
new and gaining favor for long-ish term secondary storage of
data (tape was still preferred for really long-term storage; in
lots of places, this was true even up until the 1990s and into
the early 2000s). Before that, tape dominated, with occasional
use of drums for high-speed temporary storage that was nearly
random-access. When PCs started to show up on the scene, and
started to ship with floppy disks, the name "DOS" was recycled.
Indeed, lots of early PCs had "DOS" operating systems, but these
are generally completely unrelated to one another; it was just a
common term for systems that were disk-oriented.
The MS-DOS interface, inherited from QDOS, which mimmicked that
of CP/M, has much more in common with DEC operating systems than
anything in the IBM mainframe world. The interface of IBM's
time sharing systems, like VM/CMS (now z/VM) has more in common
with Multics, or CTSS (which was the predecessor of both), than,
say, TOPS-10 or TENEX or DOS/8. It may be worth clarifying that
these things didn't usually spring forth from a bubble; a lot of
the peole who were building these things in their garages and
who started the early PC companies had some experience with
mainframe and minicomputer systems; they naturally drew some
influence from those when they started putting together the UIs
for their machines.
IBM's larger machines (what we usually associate with
"mainframes") had come out of a world that was bifurcated
between scientific and business computing; systems like the 1401
were targeted towards business, which needed high throughput,
but performed relatively simple (usually decimal or integer)
calculations. Systems like the 7094 were targeted towards
scientific computing, which needed fast floating point for
complex calculations, but relatively low throughput. To
illustrate, consider charging compound interest on a bank's
portfolio of mortgage loans at the beginning of each month,
versus calculating the trajectory of a rocket. The rules for
the former may be complex, but the math is pretty simple ("take
this number, add 10 percent to it, and store it somewhere"); the
latter is helaciously complex ("evaluate this integral to
compute the area under this curve as time varies from a to b,
but mass decreases nonlinearly as a function of fuel consumption
and decreasing drag as we move out of the atmosphere..."). A
large bank might run their mortgage interest program over a
million or more loans, while NASA's only doing the trajectory
calculation for a single mission at a time; they may run it more
than once, of course, but probably not a million times.
The IBM 360 line was supposed to unify these two worls onto a
single ISA, hence "360" in the name, as in "a 360 degree view of
the world of computing." The problem was that the software for
the 360 was famously delivered behind schedule, well after the
hardware, as recounted in Fred Brooks's masterful, "The Mythical
Man Month"; so IBM had 360 systems sitting on loading docks but
no software to go with them. While OS/360 was still being
developed, they quickly put together stopgap operating systems
so that they could move their machines into customer hands.
DOS/360 was one of those, and it was small enough that it could
run on a 360/30 with something like 8 or 16KiB of RAM and a
disk. They also shipped a TOS/360 ("tape operating system") for
systems without disks. But it was a batch system, with no real
user interface that would be meaningful in the context of a PC
or interactive timesharing system.
IBM got into the PC market largely because they saw a business
opportunity, but it's not clear that they really believed in it;
the original IBM PC project, coming out of Florida, was run very
differently than projects in New York and is a reflect of that.
- Dan C.
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
