Are there any command lines you can trace back to neither the UNIX shell or the various DOSes (or CP/M)?
The evolution of most command lines I know of go back to either the Bourne shell of the first UNIX machines (any of the *sh programs ultimately go back to sh, even csh/tcsh) or the CP/M command line (which the DOSes shamelessly ripped off and somewhat improved/extended). What I’m interested in is, essentially, an alternate route of evolution. Trouble is, I can’t think of any OSes that use a command line that can’t be described as either *nixes or DOSes.
Maybe the debug command line of the old MacOS is what I’m looking for, but I have no idea where to get information on it. I don’t even know if it fits my criteria.
So, anyone out there know of obscure CLIs? Do they even exist?
Egad, there’s a huge number of them. The ones I was most familiar with are the DEC ones: various Tenex and VMS type systems. Any interactive system, which became popular with the advent of time-sharing, are command line systems. There’s some holdover of the old DEC OSes, for example, in the way FTP works.
A lot of classic Unix utilities originated on DEC systems and were ported over. E.g., the “tar” command. If you look at the traditional way Unix “man” pages are formatted (nroff), it uses a system whose ancestry also comes from the DEC days.
In the olden days every company that made mainframe computers had their own proprietary OS - IBM, Univac, Burroughs, Computer Science Corporation(?), and DEC. I used them all at one time or another but the best one hands down was DEC’s VMS. It’s still being used but not much.
ccvm.sunysb.edu is still ticking away, and I can tn3270 into it from my Mac and run SAS (if I can remember SAS commands) or draft email in an environment without built-in word wraps if I want to reexperience my introduction to email.
VM not only has a primary command line at the bottom, it has a mini-command area to the left of each line where you can do things like copy, delete, insert, or duplicate one line at a time. Or one file at a time while viewing files.
The first “off the shelf” computer you could buy was the PDP from DEC (PDP means Programmable Data Processor, so named because they thought “computer” would make people think it was a big expensive machine from the likes of IBM or Central Data that almost no one could afford). PDP’s in their day ran RSX-11 and Unix. The PDP’s eventually evolved into the Vaxes and RSX evolved into VMS.
I know of several PDP-11’s running RSX still in use in industrial environments. They are so old that they break a lot and are a royal pain in the you know what to find parts for, but so far keeping them running is cheaper than replacing them all.
VMS is still in use in quite a few places, especially since it was ported to the DEC Alphas (the successor to the VAX). They are on their way out, but still have a very large installed base so expect to see them running for quite some time.
I know of one univac still in operation.
I also know of a wang (keep your dirty jokes to yourself) word processing system still in use. It has its own little operating system, though it’s very feature lean compared to RSX and Unix and the like.
You can get a VMS command line for NT, complete with a com file compiler (com files on vms are like bat files on a pc, only you can do a lot more in a com file than you can in a bat file). There are PDP-11 emulators out there, so if you can get a disk image of an RSX system you could run RSX in emulation mode on your PC.
I remember using Amigas back in the day that had command line shells–but I have no idea what they were based on. I remember it being pretty similar to MS-DOS, though.
Most 1980s microcomputers had the command line, editor, and programming language (usually some variant of BASIC) all rolled into one. If you wanted to list directories and run programs, you had to do it through the “immediate mode” of the BASIC interpreter. The equivalent of the Unix “ls” in CBM BASIC 2.0 was:
LOAD “$”,8 : LIST
File system commands were even trickier, since the disk operating system (DOS) was often not built into BASIC. You had to open a command channel to the disk drive and then send the command. The DOS command to erase a file was S (standing for “scratch”) followed by the drive number followed by a colon followed by the filename. To issue the command from BASIC you would have to type this monstrosity:
Amiga’s was called CLI, which was basically just a way to get into Workbench. I know of quite a few companies with some Vaxen still running. They run forever. The one I used to admin was up for 6 years without a shutdown or reboot, until a drive controller finally went.
Amiga traces its rooots back through the C64, Vic20, and PET lineage, which in some ways goes back to the CP/M command line. All of the 8 bitters (Commodore, Atari, TI, Apple, Tandy, etc) had similar command line interfaces. Most weren’t sophisticated enough to truly be compared to CP/M, but they all were very similar in design. Due to VERY limited ROM space, they were all also very stripped down and designed to function very closely with the particular architecture of the machine. You have to keep in mind that most of those “operating systems” had to fit inside an 8k ROM, which is definately not a lot of room to play with. By comparison, the boot ROM (which only contains the BIOS, not the operating system) on a PC is 128k these days.
I probably wouldn’t lump the 8 bitters into the CP/M lineage, just because their mini-operating systems didn’t really contain much. Maybe TRS-DOS and Apple-DOS, but things like the Commodore DOS worked in such different ways (LOAD “$”,8 to get a directory listing, for example) that I would lump most of the 8 bitters into their own catagory.
Not exactly. NT was put together with some major help from some ex-VMS guys, and because of that things like the task scheduler and some other parts of the OS do have quite a bit of VMS-like qualities to them (remember that VMS was doing multi-user multitasking quite well back in the 70’s so it was a pretty good OS to borrow ideas from). However, the application interface was all microsoft and bears little resemblence to anything in VMS.
I’m a little surprised by this assertion. The direct ancestor of Unix was Multics, built by Project MAC at MIT from 1961-85 or so, and I believe DEC was never part of Project MAC. The hardware was IBM and later Honeywell.
I believe nroff comes directly from runoff, which was originally written for CTSS, a project MAC timesharing OS implemented on IBM hardware in the early 60s. Runoff was the standard Multics text formatter. I used runoff to write Multics documentation as a student employee at MIT’s Information Processing Center in 1980-81. Edited the input files with a line editor on a teletype, forsooth.
At that time MIT also ran an IBM 360 with VM/SP and CMS. CMS wasn’t completely line-oriented, since you only accessed it through 3270 terminals hardwired to the mainframe, which had special keys and messages areas on the screen. But the philosophy of the commands you did give was quite different from Multics or Unix, since it was oriented around the virtual machine concept, not the idea of an abstract job. I remember to send someone a file, the predecessor to e-mail, you had to use your “virtual” punch to “punch” a “deck” of virtual cards into their virtual spool. What a hoot. Also, if you wedged your CMS, you could escape to CP, the control program, sort of like a monitor, and initial program load (IPL) CMS. Kind of like rebooting. Only it took about 2 seconds, tops.
From The UNIX Operating Environment by Kernighan & Pike:
[/quote]
The UNIX operating system started on a cast-off DEC PDP-7 at Bell Laboratories in 1969. Ken Thompson, with ideas and support from Rudd Canaday, Doug McIlroy, Joe Ossanna, and Dennis Richie, wrote a small general purpose time-sharing system comfortable enough to attract enthusiastic users and eventually enough credibility for the purchase of a larger machine – a PDP-11/20
[/quote]
MULTICS in fact has nothing to do with UNIX, other than maybe sort of in an indirect way. The name UNIX (later evolved into Unix) is a pun on MULTICS. Thompson and Richie might derived some ideas from MULTICS, but Unix is definitely not an offshot of that.
I would have to disagree strongly. First, I’ve used both systems, and Multics and Unix are obviously similar in the user experience to anyone who has sat in front of both, more so than any other two operating systems. Second, here is Ritchie on the early evolution of Unix. You’ll note he refers throughout to Multics, to how he specifically tried to salvage what was good about Multics and correct what was deficient. Third, here is an interview with Thompson in which he touches on the subject. Note in particular:
Computer.What about the development history of Unix?
Thompson.The early versions were essentially me experimenting with some Multics concepts on a PDP-7 after that project disbanded. . .
How closely related a ``direct offshoot’’ would have to be is obviously a matter of opinion. But I don’t think there can be any doubt that Multics had more influence on Unix than any other piece of software.
I just want to clarify a little on DEC OSs/Multics/Unix etc. There really was a lot of blurring. Just because tar also existed on Multics doesn’t mean it never existed on Tenex/TOPS-20, etc. Runoff, in fact, was a DEC proprietary system. But it was widely cloned. (The first “word processing” software I ever used was a Runoff clone running on a Xerox Sigma computer, itself a clone of an IBM/360.) People at MIT and elsewhere created a lot of software for the various systems available. Since most people didn’t use Multics on their DEC computers (surprise, surprise), it was the DEC OS based versions of utilities that spread/evolved/got used more. And thus had a more pronounced legacy. Look up the history of DECUS somewhere.
Keep in mind that Unix originally wasn’t a fully multi-programmed system. (It even says “Uni” in the name!) The BSD folk were the first to add in all the really good stuff that made Unix into what we think of today. They were definitely not all that influenced by Multics. (Different coast.) They were also responsible for adding in the major utilities that made it a useful system. So the main DECUS packages were among the earliest ported (well, re-written) to Unix.
Multics’ most notable legacy is the GNU OS Kernel that Stallman has been “on the verge of getting around doing any day now” for the last 20+ years.
If the DEC computers and OSs had never existed, the development of OSs would have been tremendously different. (Think Blue.)
Trouble is, I can’t think of any OSes that use a command line that can’t be described as either *nixes or DOSes.