92

In the 1980's at primary school we saw MS DOS and DR DOS competing before Windows 3.1 and subsequent releases took over that space.

At home we had Apple II's which booted up to a BASIC prompt. On these we ran PRO-DOS and CP/M.

30 years later my kids hardly know what a disk is. That made me think "Why was it called a 'disk' operating system? Was it to contrast it with a 'ROM' based operating system? (You never really heard of a 'ROS').

My question is: Does "Disk Operating System" imply that there was a "non-disk" Operating System?

  • 20
    It's also relevant that the original name of what became PC-DOS was QDOS--for "quick and dirty operating system". – chrylis -on strike- Mar 5 '17 at 6:46
  • 1
    @chrylis ...which unfortunately collides with the Sinclair QL's operating system "Qdos" - name is kind of weird (and, was, to my knowledge, never explained), as the original QL didn't even have "disks" – tofro Mar 5 '17 at 11:23
  • 4
    Acorn used the term "Machine Operating System", or MOS - Wikipedia: Acorn MOS. – Andrew Morton Mar 6 '17 at 9:42
  • 7
    Note that what M$ called "DOS" was commonly referred to as a "monitor" on big iron systems of the era. A "monitor" was a piece of software which was just smart enough to take a command, load the identified program from disk or tape, start it, then regain control when the loaded program ended. Generally it would be paired with "utilities" which provided services such as disk access, printing, etc. A true "operating system", on the other hand, had facilities for managing multiple processes, etc. – Hot Licks Mar 7 '17 at 1:11
  • 2
    TL;DR: It's more like "Disk-Operating System", than "Disk Operating-System" – Ian Mar 9 '17 at 18:19

16 Answers 16

92

The term "Disk Operating System", or commonly "DOS", was used in the early days of personal computing to distinguish operating systems that also contained software for supporting disk devices, since not all of them did. The DOS software could access blocks stored on disk, that were organized into files, and there was "filesystem" software included for managing the collection of files on the disk.

The term does not imply that the DOS software itself had to be loaded from disk. Several popular systems included the DOS software in ROM in the microcomputer, or even in the disk drive itself, as was the case with Commodore disk drives and CBM DOS. Other DOS software, like ProDOS and MS-DOS, did load from disk.

So, yes, operating systems lacking software support for disk devices would be called simply "OS" and not "DOS".

  • 6
    an example of non-disk OS is DEC Paper-Tape System for early PDP-11 models, retrotechnology.com/pdp11/11_20_PTS.html – sendmoreinfo Mar 5 '17 at 10:20
  • 1
    Right, before that, your home computer file system was 1 file per cassette tape or paper tape. Apple (Woz) figured out how to put a disk bootstrap loader on 128 bytes of ROM on the disk controller. – Harper Mar 6 '17 at 18:11
  • 4
    @Harper No; in the case of no disk operating system, or if the software simply didn't come on disks, you'd use the counter on your tape drive and an index card as a file system. – a CVn Mar 6 '17 at 20:51
  • 3
    MSX computers were one of the systems that had the OS (MSX-DOS in that case) in ROM. If the disk drive had a disk containing MSXDOS.SYS and COMMAND.COM at boot time, you were able to boot in a command line environment that was pretty much a stripped down clone of the one for MS-DOS (and which for the most part was just a thin layer on top of the ROM code). Otherwise, you booted to the MSX-BASIC interpreter (also in ROM) prompt and you were able to access disk files by using DiskBASIC (a set of special BASIC commands). – Konamiman Mar 8 '17 at 8:35
  • 2
    @Harper The PDP-8 had a loader (RIM loader) that most programmers memorized. It had 18, 12 bit instructions, and loaded programs in its own, awkward RIM format. It was used only for loading the paper tape BIN loader, which was a much better (and longer) loader. The RIM loader could have been used as a boot loader in a pinch, if the little fanfold BIN paper tape was lost, and the boot programs were available in RIM format. I never heard of it being done. Anyhow, a 128 byte loader isn't impressive. – stretch Mar 9 '17 at 14:15
63

It doesn't imply that it's the disk operating system so much as it implies that it's the disk-operating system. You could boot an Apple II from ROM, enter and run BASIC programs, load programs from cassette, and basically do whatever an Apple II can do, but there was no way to access files on disk. Apple DOS didn't really do any of the features of a modern "operating system" (which were either built into ROM or entirely nonexistent), but what it did do was provide the routines for accessing disk files and directories, thus disk operating system. The same can be said of Atari DOS, CBM DOS, TRSDOS, etc.

Although MS-DOS isn't entirely the same, to some extent it is. The earliest IBM PCs did come with BASIC in ROM (called "IBM Personal Computer Basic" on the copyright screen, but commonly known as "Cassette Basic" to differentiate it from "Disk BASIC" which ran on DOS), and the BIOS ROM provided the basic facilities of keyboard, screen, printer, etc. So the most important service that MS-DOS added to that was the ability to read and write files from disk, and execute programs on disk. It also added a command shell (COMMAND.COM) that was the most common way of interacting with the computer, rather than BASIC. When later PC models eliminated ROM BASIC, DOS became pretty much the only way to boot and use the machine at all (unlike most other 8-bit machines).

  • 3
    Great answer. Machines like the Apple //, the 1981 IBM PC, and several 8-bit Atari boxes could boot right to BASIC. For the Apple and Atari, BASIC was effectively the command prompt. Booting from Apple or Atari DOS added disk access hooks to BASIC so you could load from and save to disk. – Columbia says Reinstate Monica Mar 5 '17 at 18:34
  • 1
    Yes, and there were also alternative DOS one could use, for example if you bought an Indus brand disk drive for your Atari, it came with disks for their version of DOS, which you could use instead of the Atari DOS. It didn't change the computer's OS or way if worked for other purposes, but the disk commands and UI were different compared to using Atari DOS (which also had a few versions). – Dronz Mar 5 '17 at 21:35
  • 2
    A big reason BASIC was in ROM is the base machine had very little RAM. There simply wasn't room in RAM for a BASIC interpreter and also any useful work. – Harper Mar 6 '17 at 18:15
  • @Harper Indeed; the original IBM 5150 could be had with 16 KiB RAM and no storage, though 64 KiB with at least one floppy disk drive was probably a more common configuration. (One or two 5.25" FDDs was an optional, though common, accessory.) Wikipedia puts the BASIC at "over 50 kB"; having all that in RAM would have left very little for user programs, and on-demand loading (while almost certainly technically possible even with the hardware of the 5150; command.com did it) would have been anything but high performance. – a CVn Mar 6 '17 at 20:56
  • 1
    @MichaelKjörling I vaguely remember the 5150 had cassette deck I/O, so every base PC had something. – Harper Mar 6 '17 at 21:01
39

The term DOS pre-dates the personal computer by a looong way: the term DOS/360 was first coined by IBM in 1964 as a new operating system for their System/360 mainframe computers, to replace TOS (tape operating system).

IBM commissioned Microsoft (at that time a garage outfit) to write PC-DOS to run on their Personal Computer, which was launched in 1981. Somehow, IBM allowed Microsoft to keep the rights to the operating system, and they started selling their own version as MS-DOS. As the IBM PC manual included a complete set of circuit diagrams, there were very soon a lot of PC-clones running MS-DOS. The rest is his history.

Microsoft didn't actually write PC-DOS: they ported something that Tim Paterson at Seattle Computer Products had put together in six weeks. Tim's working title for the project was QDOS (Quick and Dirty Operating System).

  • 5
    Don't forget that the ROM BIOS needed to be reverse-engineered. (I think the ROM BIOS assembly source code listing was available, but it was and is protected by copyright.) I recall reading an anecdote, I think it was from Compaq, that they felt they had to have missed something because their BIOS was far smaller than IBM's. – a CVn Mar 6 '17 at 21:00
  • 4
    I don't think Microsoft was a "garage outfit" by 1980: they had been selling Basic interpreters to Altair and Radio Shack (at least) for years. – Martin Argerami Mar 7 '17 at 0:29
  • 1
    @MartinArgerami: Reverse engineering would definitely get you into trouble with IBM's legal department: you had to rewrite it. The bigger companies were very keen to demonstrate that they hadn't infringed IBM's copyright by having one team write a specification and another team in a "clean room" write the code to implement the specification. As you say, it wasn't difficult: the BIOS just had to provide a standardized interface to the keyboard, screen and disk hardware (it didn't even provide serial port interfaces), and be able to load in an operating system. – JavaLatte Mar 7 '17 at 1:02
  • 1
    @MartinArgerami: OK, garage outfit is a slight exaggeration: they had six personnel at the time. – JavaLatte Mar 7 '17 at 1:11
  • 1
    @MartinArgerami Microsoft was selling Microsoft BASIC all over the personal computer industry at the time. They used the expiration of the AppleSoft license as a way to get leverage over Apple to sell BASIC into the Macintosh apple2history.org/history/ah16 – mschaef Mar 7 '17 at 22:02
17

Does "Disk Operating System" imply that there was a "non-disk" Operating System?

Yes.

A little bit of history

The earliest computers did not even have Operating Systems.

The earliest computers were mainframes that lacked any form of operating system. Each user had sole use of the machine for a scheduled period of time and would arrive at the computer with program and data, often on punched paper cards and magnetic or paper tape. The program would be loaded into the machine, and the machine would be set to work until the program completed or crashed. Programs could generally be debugged via a control panel using dials, toggle switches and panel lights.

Hard Disks arrived later.

In 1953, IBM recognized the immediate application for what it termed a "Random Access File" having high capacity and rapid random access at a relatively low cost. After considering technologies such as wire matrices, rod arrays, drums, drum arrays, etc.,the engineers at IBM's San Jose California laboratory invented the hard disk drive.

Early operating systems were loaded from:

  • A switch panel

  • Paper tape

  • Magnetic tape

  • Punched cards

or some combination of the above.

Further reading

  • Even the Wikipedia article fails to note that they were simply called "disk drives" until after the development of the flexible media "floppy" drive. – gbarry Jul 7 '18 at 20:08
17

In the personal/home computer space, the usual model was that the machine shipped with built in BASIC and the ability to save and load data from an attached cassette tape. This includes the original IBM PC, which had BASIC in ROM, as well as a specific cassette port that included a relay for controlling the cassette motor. BASIC itself included a MOTOR command for controlling this relay, and enterprising hardware hackers could use this port to control things besides cassette motors. In these early systems, the OS was effectively the BASIC interpreter.

From here, the PC quickly diverged from other small computers. For an Apple ][, DOS was mainly the disk drive support software and some additional commands for the BASIC interpreter. A BASIC REPL (Read-Evaluate-Print loop) therefore served as what'd we'd consider today to be a terminal or command window, and BASIC maintained primacy over the system unless it was replaced by something that it loaded.

For PCs, however, DOS was a larger and more self-contained entity. PC/MS DOS had its own command processor and could be useful independent of BASIC. For disk based PC machines, BASIC was shipped with DOS and was a program to be run under DOS like any other. (Cassette BASIC and the Cassette port fell out of favor almost immediately, and most machines did not include either.)

So the tl;dr is that non-disk operating systems tended to be BASIC interpreters that allowed mass storage only via tapes.

Much of this was true in the mainframe and minicomputer space as well. Tapes predated disks, so pre-disk operating systems all tended to be tape based.

8

Small single-user computers were typically a single task. They also had limited memory, so the most that their operating systems could do was manage I/O. As most of their function was dedicated to disk filing and maybe a little terminal/printer I/O, they were primarily disk operating systems.

Large computer systems, on the other hand, had functions to manage multiple users, perform memory management, job accounting and network functions. Disk I/O was a relatively small and low-level part of their operating systems. Consequently, a DOS was a minimal shell over the small machine's hardware I/O, whereas a full OS was much larger and abstracted away from hardware.

  • Do you have any references for your claim? The existence of the DOS/360 disk operating system for IBM mainframes (mentioned by JavaLatte) which predates most other disk operating systems by more than a decade, seems to contradict this claim. – Dubu Mar 6 '17 at 12:04
  • 1
    Your DOS/360 link is a pretty good reference. From the History section: “DOS was designed to use little memory, and could run on 16 kB machines, a configuration available on the low-end S/360 model 30. Unlike OS/360, DOS/360 was initially a single-job system which did not support multitasking”. So DOS/360 was intended as a small, interim, single-user system until OS/360 was released. The analogy isn't foolproof, though: DEC's OS-8 for the PDP-8 feels like a DOS, not an OS. – scruss Mar 6 '17 at 14:15
8

I think it is fair to say that Disk Operating System implies an OS that talks to disk in the commonly perceived environment of files, file sizes, directory(ies), time stamps, open(), close(), read(), write(), etc. Operating System is neutral on the subject of whether or not there is a disk. Many might group them with devices that have no large, native storage, such as many classic PDAs, or that used cassette as the primary storage medium. As the question implied, we have sort of moved beyond the idea of DOS. Yes, Windows, OS X, and Linux can all talk to disk (HDD and/or SSD), but in the networked world, that is a feature we no longer really think about.

But something else exists. Classic FORTH broke all the perceived rules. Most implementations had access to disk. But no one referred to it as a DOS. Heck, most people would not even recognize it as an OS, even if it allowed a full, interactive, multi-user, (co-operative) multi-tasking environment in 64 kbytes of RAM with an 8-bit processor. Even if the OS was indistinguishable from the language. Or the application(s). Or the development environment. You had the hardware and you had FORTH, and that's all there was. If you were operating single-user, you had 100% control of everything, all the time.

The thing is, even without files, directories, or time stamps, it was an OS that gave R/W access to disk. At a high level, just four FORTH words (keywords, functions, whatever term you favor) were used to interact with the disk

  1. BLOCK (n .. addr): given the block number n, read it into memory (if not already present in one of the available buffers) and return the address of the first byte.
  2. UPDATE : mark the most recently used BLOCK as needing to be written back to disk before it is over-written
  3. SAVE-BUFFERS : immediately write all updated blocks (if any) back to disk
  4. FLUSH : perform SAVE-BUFFERS and force re-read of any block, even if it is still present in memory

There were other words too, both higher and lower level, but those were at the heart of this operating system thingy that talked to disks.

I'll add that most modern FORTHs cast away their operating system nature and operate as simply one application among many under the "real" OS.

5

"In the 1980's at primary school we saw MS DOS and DR DOS" -- in the '80's? Hah! Youngling!

Before DOS there was TOS -- the Tape Operating System. You would mount a substantial looking tape on a tape drive the size of a refrigerator, select it as the boot device using dial switches, and press the blue LOAD button.

Viola! the tape would run, and the operating system would be loaded.

  • Awesome! Love the spirit of the answer! – hawkeye Mar 6 '17 at 3:34
  • 19
    I think you meant to say voila. Violas require a language environment capable of handling mutable strings. – Columbia says Reinstate Monica Mar 6 '17 at 11:42
  • 1
    @RobertColumbia And just what "environment" is that? Does it have an OS? – Gypsy Spellweaver Mar 10 '17 at 3:30
3

Just to clarify in a single sentence: some of these systems did not have disks.

Early home computers (e.g. ZX Spectrum, BBC Micro) ran entirely from ROM and could (slowly) load programs from audio cassette tape. BBC Micros could support disks, which involved adding a ROM. If you did this it got listed on the boot screen as "Acorn DFS" (Disk Filing System) or similar:

enter image description here

2

It appears that IBM was first, but there were others. Back in the day (mid 1970s) the PDP-11 had a single-user DOS whose only purpose was to execute a sysgen program and an assembler and linker, all of which generated a tailored operating system, which could be one of RT-11, RSX-11, or RSTS/E, depending on your requirements. [I'm not speaking of the Unix sysgen process here as I never saw it at the time.]

I think there was (& is) a strong implication of 'single-user' with the name DOS.

2

Just to add to what already have been said here.

You need to remember that until late '60s there were no disks of any kind used in computer storage or otherwise.

All computing systems where using punch-cards or tapes for any data storage and input/output.

They were nothing more than huge electronic devices with what you would call active memory setup, all actions where done in memory. When you start the "computer" it would run a set of instructions stored in the read only memory module EPROM/BIOS that would check all the hardware connected to the system and would wait for the user to start some kind of action. The action would be a command to load the program from a data medium. i.e. load a punch-card from a reader, or load a tape.

Keep in mind that data storage medium progression where a board with a physical on/off switches ==> to a paper punch-card ==> to a paper punch tape ==> to magnetic tape similar to what been used in audio or video tapes.

All of these storage mediums have one thing in common, they were/are a sequential read devices. You have to start and read all of it from beginning to end. And they way it would work is that you would first load the OS in the form of interpreter and compiler, then you would load the program which would be compiled and loaded into memory, then you wold load the data which would be processed by the program and the output would be dumped to printer, or, if further processing where needed, to the input medium like punch-card or tape to be loaded later.

With the advancement of magnetic media, IBM in search of cheap, reliable, and portable storage media, developed the first floppy disk. Yet there were no operating system to properly use it. The disk was 8" in diameter, read only, and essentially was still used as a tape, until in 1974 Gary Kildall of Digital Research, Inc. Came up with CP/M (originally standing for Control Program/Monitor and later Control Program for Microcomputers) that initially where targeting an Intel 8080/85-based 8 bit microcomputers systems.

The key point of the CP/M and why it was so successful at the time was that it introduced different use of the disk media i.e. a non sequential read and later write of the data to/from the media. It had small memory footprint (only 64Kb) it was fast and reliable.

The CP/M system where a de facto standard for computing system that where using disk storage medium until 1981, when IBM ventured into Personal computing. They did initially approach Digital Research for license of CP/M to be used on IBM/PC but were not successful in that endeavor and went back to Microsoft which in turn came up with PC-DOS and later with MS-DOS.

  • WOW, haven't noticed so many misspells. Thanks JAL. – vlad Mar 10 '17 at 15:41
1

The Tatung Einstein was a mid-80s 8-bit home machine. When booted without a disk in the drive it would start a monitor interface (hex editor and debugger) called MOS - the Machine Operating System. This interface could then chainload DOS from the floppy drive. The DOS it used was a CP/M deriviative.

1

In the mid to late 1960s Digital Equipment Corporation's 36-bit PDP6, and its successor the 36-bit PDP-10, had a DECtape-based operating system called the Monitor. It was a multi-green-screen timesharing system. DECtape was a highly-reliable magnetic tape system providing random access to numbered blocks of data. This was later superseded by a disk-based operating system, but prior to that in 1969 custom disk access for data storage was developed for the Melbourne Stock Exchange in Australia.

0

As I recall, early desktop computers (what became known as PCs, as well as the first Macs) didn't have hard disks, only memory (ROM, RAM) and, later, removable floppies. The computer's BIOS (Basic Input/Output System) was the firmware that booted the computer. It had to be loaded first to get the computer up and running before any DOS could be used.

0

Back in the day, we used a Data General Nova computer with a 9-track tape drive. It had non-volatile core memory. And when you use RTOS (The Real-Time Tape Operating System, the 9-track was used much like hard drives are used now. There was a directory at the beginning that would tell the tape unit where to find the files. We used it for scientific research, and stored both data and programs on the tapes.

No disks on that system at that time.

0

I learned programming on an IBM 1620 (6-bit BCD magnetic core memory) computer in 1965, and it had an attached IBM 1311 disk pack drive and IBM 1622 punch-card unit. We had no operating system available at the time, and wrote everything in raw machine code or assembler. There was a way to load and run the assembler from disk, but I don't recall the details...

  • 1
    Welcome to Retrocomputing Stack Exchange. Please read the tour. This doesn't strictly answer the question, as the question is asking whether there was a "non-disk" operating system and you are talking about a system without an operating system. – wizzwizz4 Aug 12 '17 at 7:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.