Whence Operating Systems?

[ChrisGreaves]

I attended a NaSPA (North American Systems Professionals Assocuiation) meeting last night.
Kellman Meghu spoke on Security in Virtual systems.

An interesting point arose:

We think of operating systems as essential - probably 95% of users on the planet use Windows or Mac or a similar device; cubicle farms used to have people sitting at OS/360 terminals and so on, but Kellman pointed out that as virtualization progressed, the need for operating systems, as we know them, largely disappears.

In the crudest sense, if you don't have ethernet cables connecting virtual computers, then you don't need O/S code to check that they are present and functioning correctly; the data transfer takes place within the virtualization layer.

The IBM 1620 (1967) had no need of an operating system; I booked 4 hours on it, fed in punched cards and squinted at the results.
The IBM 1401 (1968) had operators in white jackets who mounted write-protect rings on magnetic tapes, and decided whether my job was important enough to be run today, or whether it could be left to be ignored by the night shift.
The ICL 1900 (1970) had an operating system GEORGE III which scheduled jobs, disk space and, amongst other things, supported virtual paper-tape devices, although we never came withing 8,000 miles of anyone using genuine, real, tearable paper tape.

In the crudest sense, if you don't have paper tape devices, then you don't need O/S code to simulate paper tape devices on disk; you store data as "disk files".

In 2010 I am almost back to where I started; I am the sole user of this (laptop) computer, and yet Windows XP sits between me and my word-processing software.

I am fascinated by the view that Operating Systems might be seen in retrospect as a fifty year temporary measure, much as we can view slide rules today.

[John Gray]

I really hope that your Mr Magoo received a robust response to the specious twaddle he appears to have been talking!

The IBM 1620 (1967) had no need of an operating system...

With Professor C E M Joad, I say, "It all depends what you mean by an operating system"...
The IBM 1620, and those IBM computers before and since use microcode ~= firmware to control the hardware, and implemented an instruction set in its assembly language, and whether you call this an operating system or not depends on your point of view. No Fortran programmer would have thought that the high-level instructions contained on those punched cards of yours were actually acted on natively, without being interpreted by the compiler (perhaps into assembly code), and thence fulfilled by the microcode!

In the crudest sense, if you don't have paper tape devices, then you don't need O/S code to simulate paper tape devices on disk; you store data as "disk files".

And you think the disk files just get 'written' without requiring any code to do put the record in the correct place, perform error checking and correction, and so on? You never did any channel programming for IBM 3330 DASD, for example?

I am the sole user of this (laptop) computer, and yet Windows XP sits between me and my word-processing software.

And the word-processing software could of itself read and write files from/to disk? And drive a printer?

There are many layers and levels of software between a human and the 'end results' from a computer: firmware, drivers, operating system code, application program code, and so on. The boundaries between these are likely to change with time, but the data transfer and control functions they implement will still be needed in some form - even if/when we don't know they're there...

[PS Was your post intended as a wind-up?!]

[StuartR]

...[PS Was your post intended as a wind-up?!]

John, Thanks for the comprehensive reply. I thought about responding to this post but words simply failed me.

[ChrisGreaves]

[quote="John Gray"][/quote]John,

Thanks for the response, and my apologies for a late follow-up.


JG> The IBM 1620 (1967) had no need of an operating system... With Professor C E M Joad, I say, "It all depends what you mean by an operating system"...
Well, quite so. I might have been more definitive. The operative syllables were “operator”, loosely defined as the-guys-in-white-coats.

To that end, I was the operator of the 1620, and (in retrospect) occupied all 3 positions of operator, applications developer, and end-user.
In that sense I maintain that the 1620 had an operator – me – but it had no mechanized system for scheduling equipment such as tape drives or disk space. Or processor time.

JG> The IBM 1620, and those IBM computers before and since use microcode ~=
firmware to control the hardware, and implemented an instruction set in its assembly language, and whether you call this an operating system or not depends on your point of view.

Thanks. I am aware of microcode from at least three vantage points – but I was not aware of it back then We hade to design computer circuits from the basic NAND/NOR gates available in a DEC handbook back in the early 70s. I studied Ken Iverson’s “A Programming Language” – again in the early 70s, and built myself a single-instruction computer out of software, same period.

JG> And you think the disk files just get 'written' without requiring any code to do put the record in the correct place, perform error checking and correction, and so on? You never did any channel programming for IBM 3330 DASD, for example?

Not at all; we all were aware that “things went on” below the level of our FORTRAN/COBOL junk; we all knew about and had written in assembly language, and knew there was something deeper. It wasn’t until the years 73-74 that I cracked open the books on Nand/Nor gates. I recall in 1968 that the 1401 had no operating system in terms of scheduling tapes or CPU; tapes were handled by the-guys-in-white-coats and the i401 was a single-program device – where program is defined as an application coded into a loadable deck of object code on punched cards. I recall programming the IOCS for a lark, mainly switching the assignment of tape drives so that the the-guys-in-white-coats wouldn’t have to physically take a reel off drive 1 and replace it onto drive 2. I saw that as a waste of time.

JG> There are many layers and levels of software between a human and the 'end results' from a computer: firmware, drivers, operating system code, application program code, and so on. The boundaries between these are likely to change with time, but the data transfer and control functions they implement will still be needed in some form - even if/when we don't know they're there...

This is true; I try to remind students that Windows can do a better job (than a human) of swapping stuff out of RAM if only because Windows sees the program not as ‘Word” or “excel”, but as a collection of chunks of memory.

To users of 40 years ago, an operating system was the visible, tangible (in the sense of JCL cards) layer that stood between us and the results of our programs. The-guys-in-white-coats were a layer that stood between us and the operating system. Our job deck didn’t get contemplated by the operating system unless the-guys-in-white-coats cared to drop it into the hopper.

Today from most computer users points of view (I’m speaking of folks sitting at a keyboard, at home or in the office), there is an operating system but no guys-in-white-coats.

For the majority of the western population, there is no consideration of an operating system at all, including, but not limited to, those who use ATMs, those who use the online catalogue in the local library etc.

I see it as a good move, and Kellman Meghu’s point was well-taken. I make my money in the computing area from businesses that run Windows machines, but the truth is I couldn’t care less about Windows. “Give me a VBA interpreter and I’m happy” might be my credo.