Two further details - XENIX - like UNIX - was a virtual memory capable OS. This gave it more ability than a 640 Kilobyte limited MSDOS. Second - and unrelated - some viewers may not know but hard drives of the early 1980's (when the 5 1/4" form factor (same as floppy physical size) first came out did not have perfect magnetic surfaces. So drive failure was a common event and you had to format the full drive (usually 5 to 30 megabytes) and record the "bad spots" to avoid using them. Put that in perspective - a single PDF file with images can easily exceed that drive capacity.
It was VM capible, if the hardware supported it, which the 8088 version did not, it was a swap file situation, and the 286 version, that Unisys figured out how to put VM in DOS, MS still did not, it was not until the 386 versions that they supported VM. The earlier 8086 version supported VM with special hardware. Even though the drives I had ( a perfect ST-225, and a 2 error CMI ), I still turned off error correction and took 24 hours to test each drive ).
The FAT had an extra flag for defect sectors and turning on the VERIFY-feature (read cycle after writing, halving the usable speed of the disk) made sense, before IDE came out. On the other hand, the disk parameters were not fixed, but set in the BIOS, and people succeeded in over-formatting their e.g. 40MB MFM drive to 60MB, by using more dense cylinders and sectors, and speed up by testing the interleave factor.... these drives were more analogue tech, than digital.
@@martinb.770 Making denser cylinder and sectors increased the rate of bad sectors, which reduced the usable size, so it hardly ever was a perfect solution, and it also increased the corruption rate of the drive. You had to find a good balance, which usually was the manufacturer suggested values.
@@Bubu567 Sure, but it was possible and as HDDs were insanely expensive, some people tried and did test the limits. Just the same happened at the end of the floppy era, when special tools pushed the 1440kB beyond 1720/1760kB, if compatibility to other drives was not needed.
Yup! That’s how we used tar for backups back in the day-direct to block devices like tape drives. This explains why a -f parameter is necessary to specify a file instead of a block device. Great stuff! Thanks for doing all of the work to bring us this walk down memory lane.
I worked at a Radio Shack store around 1993 or so, and the POS terminals out front were connected to a Xenix system in the back room that took care of all of that stuff. :)
Thanks! Glad you like it. Though no more SGIs for the time, I'm afraid. I'm currently in the mid-90s DOS/Windows period, and will soon do some stuff on networking. Hope you're going to like that as well ^^
@@THEPHINTAGECOLLECTOR Don't worry, as a fan of technology as a whole, I'm always interested in finding out about new things, so I'll be looking forward to those as well 😄
Back in the '80s I helped to administer an AmDOS (American Dental Office Systems) implementation which ran on Xenix. It was fun, and I learned a ton, since it required diligent care and feeding. Thank you for bringing back memories.
Oh wow blast from the past! Used to support Xenix and SCO UNIX at small businesses that had serial Wyse terminals and serial printers. Later discovered Slackware ( lots of floppy disks ) as first foray into Linux.
Same. We had Stallion serial multiplexors for serial for printers / dumb / "smart" terminals. Started with something called Yggdrasil Linux/GNU/X, or LGX, then Slackware etc. Then into Debian / Redhat.
Hehe you mentioned SCO, I followed the litigation soap in the naughties, Darl McBride was one slimy SOB. I've never seen a court case last so long without the accuser ever handing over concrete proof of plagiarism. They insisted linux had copy/pasted bits of code from SCO UNIX in it. In the end, SCO had nothing, they were just trying to cash in on Linux. Probably the most expensive chain of litigations ever, several subsidiaries pumped hundreds of millions into SCO, one of them being Microsoft - to this day I can't fathom why, but maybe you can shed more light on that?
@@joeturner7959 Its been too long :) Looking at pictures though, they looked more like the Wyse 50 terminals with monochrome displays (I want to say green, but not 100% sure).
To this day I still have a SCO5 server running, Xenix was a little before my time but I supported a customer or two with them early in my career. A most enjoyable video... I have a feeling that it is just delete and not ctrl-alt-del that will get you out of Snake, as del is the universal "get me the heck out of here" key on later versions of SCO Unix.
It's surprising how fun is playing with those unix systems and how much of the legacy commands as "message" is still included and working on modern linux systems
I too, argue with the Mr. Know-It-All in my brain. He usually wins, you're better at knowing things than me, clearly. Smashed that sub button, great channel sir!
I worked with Xenix with some of our clients on Intel. It was surprisingly solid as an OS. Typically they were migrated to SCO UNIX. Maybe not so surprisingly as it was a licensed copy of AT&T Corp's UNIX by Microsoft. SCO OpenServer V was very stable and reliable, but lacked a lot of the latest Linux innovations like a decent shell, so you had to port it across or use Skunkware etc. Compared to Linux with some items, it was very slow with certain actions. We had some DEC Alpha servers, but they were on their way out at the time, extremely nice hardware though.
Xenix and other older Unix versions use UUCP with getty and a modem. They send around files via USENET with UUCP. I don't believe slip and PPP were even a thing back then. UUCP did everything.
UUCP could distribute USENET-posts, and those posts often included files encoded with the UUCP-command UUENCODE. But USENET really had nothing to do with UUCP. UUCP doesn't need USENET. It's a kinda neat protocol, I use it daily. There's also IP implementations of it.
8 MiB in 1983 was... unfathomably huge outside of the biggest corporate and university IT departments! My first HDD was only 5 MiB and that was a mind boggling amount of space at the time.
By 91 to 93 I worked with Xenix. There was two servers (at the place I worked to). One was a 386-20, with 8MB RAM and 16(!) serial terminals hanging to it. The other one was a 486-66 with hopping 16GB RAM with more 16 serial terminals. They ran a extended version of DBaseII (Clipper like) called TSGBD. Never had an issue with the OS.
I worked for a wholesaler that ran entirely on Xenix up until at least the early 2000's when I left there. Also had a fairly impressive old line printer that was insanely fast and would still have jobs backed up at times. Fortunately I was the sole nerd there and knew how to manipulate the print queue to make me the next job. Led to some confusion, I just played it off like I was dumb about that stuff as they were.
Also, With SCO Xenix and Unix, installing a dos partition on the Hard Drive, and installing MS-DOS on it, and then Xenix / Unix on the rest of the drive, allowed you to copy files between the Dos partition and Unix partition with commands in Xenix / Unix. In Fact, you can boot into either DOS or unix using the boot loaded that comes with Xenix / Unix.
I have seen these in action. You need a serial card in order to network it (the name of the card type still escapees my mind). Well, network in the guise of having serial dumb terminals, as I don't think it hadsny ethernet nor IP connectivity until much later.
@@joeturner7959 As far as I can tell that's not correct...? I think it's like this... 1.0: Unix version 7 (no "System") 2.0: Unix version 7 (no "System") 3.0: Unix System III 5.0: Unix System V release 2 (SVR2)
Xenix was used with Ungermann-Bass Ethernet cards (with twisted-pair adapters) and the XNS protocol (unswitched) for corporate LAN (along with IPX/SPX and netbeui). The xenix machines were used as mail servers with Microsoft Mail client software until exchange and outlook replaced them. There was even an internal port of the xenix mail server to use TCP/IP on NT 3.1 so that the LAN could be segmented (exchange was delayed)
Ctrl+Backspace can be used on later versions instead of Ctrl+C. PErhaps here too? You could try stty intr "^C" and see if that works for changing to ctrl+c
Any Z8000 Xenix systems around? That would have been my system of choice back in 1978, though the 68000 had more possibilities. In the end, around 1980, I got a Wicat 68000 with its own DOS-like OS. Chrislin 512K memory board which didn't work unless I blocked off the top 48K.
the TCP/IP tools are a layer on top of what ever networking you use to connect the machine. Its an era thing, T-base 10 (coaxial) was still a new thing with machines in the 90's, where slip was still very common, ptp (or ppp?) not so much
Floppy problems aside, is there any chance to run a downloaded Xenix on "modern" hardware today? Mostly interested in having a working pre-K&R C compiler,
There was a version called Xenix 386, which propably works on later hardware up and until the early 2000s. Maybe not the present-days hardware, as it doesn't know anything about SATA and other modern days hardware. Would be worth a try.
Of course there is. Use a virtual machine like virtualbox or QEMU or a real emulator like Bochs or PCem. This is your only option on very new computers as they no longer have UEFI Legacy BIOS support. In addition, the A20 gate support on modern Intel mainboard chipsets is missing too, but that shouldn't be that important for Xenix.
slip and a 1200bd modem is where i started … .. felt kinda magic at the time. Tar was The most incredible swiss-army knife at the time… especially when I later got some Sun workstation with a tape drive .. … too bad I gave those away… but it was fun to run linux/x11 on them in the 2000s …
Back to the 80's, the best operating system from an architectural point of view was QNX 2.x. Multi task, multi user and realtime ! There was no Ethernet networking (TCP/IP), but setting up an Arcnet network was a piece of cake and you could easy access all the ressources from a remote machine as you would for your local machine. An OS ahead of its time !!
i used to use SCO XENIX in the late 80's early 90's... if you have a dual operating system you can type dos at the xenix boot prompt and it should boot into the dos partition without having to swap active partitions with fdisk.
i cut my teeth on Xenix way back in the 386 days. they were used to run major call centers alongside serial terminals, message boards and sat hanging off the PBX switch. I still have the SCO cards with the license numbers on them and still have a couple of VMs running Xenix, Unixware and openserver with their respective dev kits
@anthonygelona9479 You would make a good MR KNOW-IT-ALL as well ;-) Honestly? I had to make a break to get the video out. There's so much stuff in there to explore, too much after all! :-)
I've played around with Xenix on the Apple Lisa some. The micnet networking was interesting a friend and I got that set up between a Lisa and an IBM XT.
XENIX was my first Unix experience. I know it was influential and Microsoft did not get enough credit for its pioneering. When SCO Unix acquired it I wondered if their marketing should have included "We sell C shells by the seashore." ;)
We had about 400 Xenix servers running along with about 300 Netware boxes. They were very reliable boxes.... Would reliably crash daily. The entire air-services radars in Australia used Xenix - and was quite good for the time. The Netware products were significantly better for file and print. We ended up migrating to Linux over a few months. Much more reliable.
Used to store raw tar files on floppies instead of filesystems, in addition to splitting them with dd. It's was so much faster & more efficient for large files.
@Pinguinomannaro74 I'm not so much into desktop systems, for the most I collect portable systems. There were though IBM portables from the PS/2 era, notably the P70 and P75. I'm looking around to get hold of one of these since a long time. And the day I get one, there will be also a video about it, promised.
At these times it was not uncommon to write tar, cpio or dd directly to the external media. If I remember correctly Interactive Linux was delivered that way ....
Amazing to watch in some aspects. One side of the Unix medal screams about complexity of Unix for weak 8086-not-so-weak-286, another point why MS ditched Unix, they felt like creating something better and simpler. Dos was fast, small and glory to that marketing department who envisioned easy tools for programmers and lots of software for users.
DOS and Xenix were meant to compliment each others, they existed side by side and catered to different needs. It was co-operation with IBM on OS/2 as well as Windows (which originally were meant to coexist as well - with very similar idea). DOS and Xenix were the old plan - OS/2 and Windows were the new plan meant to replace them. And really, DOS lasted way longer than it had any justification for doing so ;)
With SCO Unix and its tcp/ip package, that package needs a key as well and you couldnt reuse the same key on systems that were on the same network ....
If you are not supporting large binaries, you can shrink the swap area to 1280, ( 640k/512b block sizes). If running large binaries, try a larger hard disk ( My hard disks were 33Mb, and 31Mb, with a 2560 swap size on its own partition. It did not help ). ( you also forgot to empty the /temp folder of the tar files. For TCP/IP you need a 386. For GNU Tools, and a decent compiler, NOT The MS 'C' Garbage that MS got from Lifeboat., you need a 386. The disk images are up on pcjs for the dev tools, and I used them to get through my 'C' class, and pass, but... The teacher started with The Lifeboat compiler, and a student complained about something... the teacher sat down at the students computer, and typed a simple program, and compiled it. He looked at it again, and then picked up the box for lifeboat C, and threw it across the room at the garbage can, and announced that we will be using another compiler, next week.
It's most certainly "networkable" - I recall seeing a mention and possibly even screenshot of Internet Explorer on Xenix; Don't ask me which version, must have been really early... I've always been interested on whether the Xenix IE port was 16-bit and able to run on 286 CPU's, or if it required the 386 version? What I was most interested about, and was bummed out because the video didn't cover any of it, was the software implementation for things like multitasking, swapping, etc. - like did it use 286's ability to do pre-emptive multitasking or did it rely on co-operative multitasking; and what about on pre-286 CPU's, was there any way around having to use co-operative multitasking and let the application steal the control by simply running as long as it wants without returning the control back to the OS? And what about swap space - I notice that it asks you for it, but on a hardware without virtual memory support that would allow swapping to disk, so is that swap space going to be just waste of disk space, or did they implement some crazy scheme of "co-operative memory management", where the application can tell the OS that "this memory can be swapped out" and then everytime it wants to use it, it has to call the OS to make sure that the stuff it wants is in fact in RAM? I mean, that's just a crazy example of what I was thinking as solution for just-for-fun OS project that would run on ancient PC's, I just mentioned it as example, but whatever Xenix actually did or didn't is what I'm interested about :P I'm assuming, partly supported by comments I already glanced at, that the swap space wasn't useful on pre-386 hardware, but then why would this version of Xenix ask for it? Or does it include support for 386 and operate in protected mode if available - and if so, I don't think the user space applications included had two separate versions, so then would it run them in Virtual x86 mode and the applications be just 16-bit real mode code? Oh man, I could just go on asking questions and speculating - it's a really bad habit of mine, this sort of questions are better asked one at time, so you can get an answer and don't have to ask questions based on erroneous guesses about that first question... Like if I just waited to find out if it even used the swap space on pre-386 CPU's, then I wouldn't have to ask speculative questions of how did they implement support for swap on 286's or even earlier if the answer was that they didn't :D Still, thank you - this is the only video I've seen about Xenix specifically, I've only seen it mentioned briefly as UNIX system for early x86's without it even being shown in use (and not just some old marketing screenshots).
During the 80ies / around DOS 3.3, it was still speculated, that MS will enhance DOS to being Unix-compliant by DOS 5.0. Needless to say, it never happened, and even POSIX-compliance took until NT 3.51, in the mid-90ies.
although you might have to construct the file lists by hand, you should be able to down grade the 1.2MB install disks to 360KB disks, but you will need the "tar disk format" tool to do it properly - that should be present somewhere, but may also only be on the Developer Tools ... before doing that, I suggest temporarily hooking up a 1.2MB drive, and seeing if your disk images are good, before wasting time on a dead end (due to bad images) - the "tar disk format" tool may already be installed with the basic (extra) Operating System install
Xenix was technical wonder back then. Single intel 286 PC can serve 8-16 concurrent terminal or telnet session user comfortably. It was good and cheap departmental office machine. We developed ditroff backend for Canon laser printer (300dpi) to print beautifully typeset document for research paper and books in 1986. Way better than CP/M WordStar word processor with Diablo daisywheel impact printer.
I think the youtube algorithm likes this video, there were 645 subscribers and 4k views when I started and 717 subs and 6.1k views when I finished. I also used Xenix and SCO3 and 5 in the 90's.
Yeah, I’m totally blown away by the attention this video receives. In relation to my small niche channel, views and subscribers are totally skyrocketing! Seems like everyone is into Xenix :)
not sure if this was said already, but it bears mentioning that xenix was a real-time OS, unlike most 'nix's of the time. we used it for control systems and such that demanded predictable responses, when hardware was slow and the OS could get bogged down by a single poorly written app
omg you didn't even show what would be the most interesting aspects of running Unix on 16bit chip without any protected memory! did they implement this in software somehow? Is all disk I/O done through the PC BIOS? Show us the weird!
Good, but confusing. :-) The entire history of it is confusing, if we think about it. The thing could be even more confusing, considering (a) Novell actually selling of the UnixWare product, coming out of the UNIVELL joint-venture to said Caldera, later rebranded THE SCO GROUP (though essentially keeping the UNIX copyright) (b) The entire DRI (Digital Research) timeline, including CP/M and DR-DOS, eventually transitioning via Novell over to Caldera, Lineo, DeviceLogics and DR DOS Inc It's all a big huge labyrinth of acquisitions, mergers, splits, rebrandings and more over the past decades.
I'd argue that compilers still isn't included in (fairly) modern UNIX OSes. AIX which we still use a lot at my company requires a license. Solaris used to be the same the last time I used it in a serious fashion back around 2005. (You obviously had the option to use GCC but that required other libraries etc to make the executable portable between machines)
I was a bit unprecise in my wording about "included". I always ever had access to the full disk sets, which included the compilers to install. It's the same on SGI IRIX as well, it came with the development foundation and compiler CDs, and they could be installed right away. It's correct though, you still needed the extra license to use it.
@@Melechtna Such a word "Schusauen" doesn't exist in German. The English sentence "Take a look at sth." is translated to "Wirf einen Blick auf etwas." in German. a = einen look = Blick at = auf sth. = etwas. But the difference between "take" and "wirf" is, that "take" means "nimm" in German, while "wirf" means "throw" in English. Or in other words, no one says in German: "Nimm einen Blick auf etwas." Xenix is a type of Unix operating system for the x86 processor and was used in the 80s before Windows NT.
@@OpenGL4ever That would be because my phone tried correcting it several times, and it didn't like me typing Schauen, and I failed to correct IT the 6th time.
@@OpenGL4ever Also, why are we using Xenix as an example of an object to be looked at, that just seems like a really bizarre and random thing to throw in there.
I would have used an XT-IDE adapter, less expensive ;) Though luckily not needed as the drive could be revived and so the machine remains fully in original condition.
XENIX developers where maybe the only ones who managed to use the 80286 crippled memory management and protected mode implementations more or less ... Everybody else went away in horror and came back when the 80386 was released.
What about OS/2? I'm not absolutely certain if it was DOS applications that it was used for (that's what I recall), but it ran in 286 protected mode, and unlike Windows (which would just switch to fullscreen text-mode and give the control over to the DOS application and resumed when it finished - plus I don't think it ran in protected mode; I'm talking about 3.0 that we had on our 286, though if you had a 386 you could run it in "386 mode" and multitask between Windows and DOS applications), it could even multitask between real-mode DOS programs and OS/2 programs - of course Xenix didn't have to switch between real-mode and protected mode, but I'm mentioning that feature to emphasize that it ran on protected mode. If not obvious, I'm talking about 16-bit OS/2 that ran on 286. Also, I'm not 100% certain, but almost anyway, that Minix (either 1.0, 1.3 or 1.5 I believe) used 286's protected mode. There were others as well, including some less known DOS applications that enabled multi-tasking, or at least task switching, on 286 and I believe some competing DOS operating systems. But I'm moving on increasingly uncertain areas, so let's just say that OS/2 is definite proof that Xenix wasn't the only one.
@@robsku1 Windows 3.0 ran in protected mode on a 286. In Windows 3.0 this was also known as standard mode. Windows 3.0 also had two more modes: Real mode and Enhanced mode. Real mode runs on an 8086, enhanced mode requires a 386 or better. Enhanced mode took advantage of the 386 CPU's VM 8086 mode for DOS applications. Thus, you could run multiple DOS applications in their own VM. The VM instances ran in preemptive multitasking, the Win16 applications only in cooperative multitasking. I'm not sure about Win32s applications. In Windows 3.1 there was no more real mode support. And in standard mode you could only run one DOS application at the same time. To do this, Windows 3.x switched back to Real Mode and DOS. It wasn't just a fullscreen text-mode, it was DOS using DPMI, while Windows 3.x was kept mostly in extended memory. OS/2 did also run in protected mode on a 286 and was also limited to only one DOS application at the same time. The 286 lacked the VM 8086 support, which was required for more. Windows NT was Microsoft's first OS designed purely for the 32-bit protected mode of the 386 and later.
You can get more out of your hard drive if you use a RLL (run length limited) hard drive controller as old hard drives are basically just dumb devices you can cram 1.5 times more on the drive. I am not sure that is helpful. But I try to be helpful.
Seems the Internet was a little bit too late for Xenix. Berkeley Sockets wasn't developed on BSD until 4.4BSD, in 1983. Cu should be available though, with uucp. The wonders of slip and later things, you may have to compile.Unix of the 1980s was clearly a far cry from even lowly Linux, lol. How much difference 13 years makes.
A bit harshly said, but true. Unix at the time was too demanding, and didn‘t establish well outside specialized niches and the server market. And licenses were to expensive for most private users, which indeed was one of the arguments for Linus to create the Linux kernel after all. Now, Linux is everywhere. But let’s not forget that there is still a descendant of the original AT&T Unix out there in the consumer mass market, which is Apple‘s macOS with all its iOS derivates, accounting for 1.8 billion devices out there. So while the Microsoft Xenix branch ultimately was a deadend, the BSD-derived macOS (NeXTstep) branch has made it.
Nice video, but this constant interruption by your alter ego, and especially the windows closing sound at the end of the interaction, is a bit annoying.
Talking to yourself is a manifestation of a split mind. Wanting so badly that Microsoft ever has made an operating system all by themselves is impossible. Everything sold by Microsoft was made by other persons or companies. It was bought for a few pennies from the silly bastard that was not able to bring the product to the market. Unix was made by the telephone company, BASIC was made by some school. DOS was a copy of CP/M, Windows was found somewhere in Palo Alto at some research company.
"Unix was made by the telephone company": correct, by AT&T and the Bell Labs (licensed) "BASIC was made by some school": Correct, by some people at Darthmouth College (didn't find if that was licensed, bought, or whatever) "DOS was a copy of CP/M": Correct, by Seattle Computer Products (licensed, later acquired) "Windows was found somewhere in Palo Alto at some research company": Nyah, not sure here if they actually took (as in bought, licensed) code base. For sure, strongly inspired by the XEROX ALTO, found at XEROX PARC, yes. Apple did the same, btw. "Talking to yourself is a manifestation of a split mind. " - Well, no. I'm talking to an avater inspired by Microsoft Office 95's clippy, which is just totally annoying. But I seem to be getting the point about what the other guy commenting yesterday meant about it, that it disturbs people ... Well ...
@@TiagoJoaoSilva It differs a lot! It is the principle. If an idea for software is developed with good standards by people with a right mind, then the result could be something that is good and useful. If an idea is bought and then further developed by others, then that company's goal is to commercialize ideas of others. That is a totally different starting point then problem solving with the help of computers. Because there is a totally different goal, the result of both kinds of software is not the same. All software made by Microsoft needs well trained technicians to install and operate. Software for the Mac does not need technicians to work on it, the aim was to make software for creative people that can be used intuitiously.
@galier2 /usr in my understanding goes OK for "user", as that's what the Linux documentation project also calls it: "User System Resources" One may correctly argue that Linux is not old enough to claim naming it, but also the BSD-line hier(7) man page [man.freebsd.org/cgi/man.cgi?hier(7)] always ever refered to /usr as "contains the majority of user utilities and applications", and the BSD traditionally contained even the home directories beneath /usr/home. The UNIX Version 7 hier(7) as the oldest I could find, refers to /usr simply as "general-pupose directory, usually a mounted file system" [www.tuhs.org/cgi-bin/utree.pl?file=V7/usr/man/man7/hier.7] with no reference to it being "Unix System Resources". I stumbled accross this one here [unix.stackexchange.com/questions/103347/pronunciation-for-usr-directory], I found this claim: "In this 1982 film from AT&T that features the UNIX inventors, The UNIX Operating System, Brian Kernighan pronounces /usr as "user", around the 13:41 mark" Now, honestly, I don't know. It may well be that it originally was intended to be "Unix System Resources" some 60 years ago, though the only notation I personally ever had been confronted with in 25 years, was the notion above about the directory containing user-related data and binaries. According to my not very indepth research, there is some controversy around it.
I've heard /usr been referred to as "usser" (not "user") from several poeple. I'm not sure if they have invented that by themselves, or if it's common elsewhere.
it's not really 'long forgotten'. it's basically multiplayer dos or cp/m and as such most likely parts of it made it into the os/2 kernel and are now simply called 'windows 11'. with some fancy new gui and new colors. it's just reselling the same shit over and over again under different names. it is also not really quite a 'unix unix'... more in the league of other programs that have dos/cpm api calls and can run multitasking, such as that quarterdeck stuff. xenix has barely any networking capability for one thing. but hey. os2 which later became windows nt 3 had to come from somewhere no? LOL. this is not something that is in the leauge of things like sys-v or dec vax. this is more like a glorified task swapper.
I strongly disagree here. The early releases of Xenix were definitely a direct evolution from licensed AT&T UNIX Version 7 source code. Still, it has nothing to do with either of the DOS, CP/M, and later OS/2 or Windows NT line of operating systems. Attributing Xenix of beeing a "dos/cp [..] task switcher [..] like quarterdeck" is inadequate. It is correct, that DOS was essentially a clone of CP/M, of which naturally a lot was inherited into the DOS VDM present in both OS/2 and Windows NT. However, the OS/2 kernel clearly is neither DOS, nor CP/M, and not of UNIX origin either. Windows NT has indeed taken some inspiration from OS/2 and even more from VMS, but again is neither a DOS, CP/M nor UNIX descendant. And if you watched closely, the SCO Xenix release 2.1.3 showcased in this video reports itself as UNIX System V, which originally came out in 1983. So Microsoft and SCO definitely had updated their source base from UNIX Version 7 to to System V at some point. It is also well-known that Xenix System III as being an intermediate release, was derived from AT&T UNIX System III. I didn't find the earliest Xenix release to play with, but having one would definitely exhibit an even more archaic look and feel than the System V release we've seen here.
@@fnjesusfreak then why is it's entire range of int21 functions essentially the same as cp/m and dos... :P i'm pretty sure no actual unix ever had that :P
ofcourse one can damn well make that reside on top of the task scheduler borrowed from any actual unix... but ya know. it's still just multiplayer dos :P
„Xenix is a discontinued version of the Unix operating system for various microcomputer platforms, licensed by Microsoft from AT&T Corporation in the late 1970s. The Santa Cruz Operation (SCO) later acquired exclusive rights to the software [..]“ Source: en.wikipedia.org/wiki/Xenix It is explicitely stated in the video, that Microsoft licensed UNIX from AT&T. It is correct, that Microsoft didn‘t invent UNIX, and no such claim is made! But still, Xenix was Microsoft‘s Unix version for the time, though they couldn‘t use the UNIX name, as they only licensed the source code, hence they named it Xenix. If anything, then maybe you should check the facts as well.
@@Theineluctable_SOME_CANT i wasn‘t there, but all sources claim otherwise… en.wikipedia.org/wiki/Santa_Cruz_Operation#Relations_with_Microsoft „Xenix is a Seventh Edition Unix-based version of the operating system that Microsoft worked on, initially for the PDP-11.[3] SCO first began working with Xenix in 1981.[8] In 1982, Microsoft and SCO forged a joint agreement for development and technology exchange, with the two companies' engineers working together on improvements to Xenix.“ „In any case, intellectual property rights were not transferred in the 1989 agreement and SCO would continue to pay Microsoft royalties for Xenix and Unix technologies.“ „In 1983, SCO made a technically difficult port of Xenix to the unmapped Intel 8086 processor (earlier 8086 Xenix ports required an off-chip MMU) and licensed rights from Microsoft to be able to ship its packaged Unix system, Xenix, for the IBM PC XT.“ To me, this does not read like „xenix was SCO‘s“ but more like „it was Microsoft‘s“. For sure SCO contributed to the code base, and for sure, they later acquired Xenix from Microsoft. But wether SCO was there from the beginning on or not, I find no claim it was their original product. Also, the particular version of Xenix System V v2.1.3 shown in this video shows the Microsoft copyright clearly predating SCO‘s copyrights.
@user-sf5bz5nh9e "Being there" implying being a SCO employee, so having first hands hard fact information? But then please, proof it. Tell me, why all sources tell different. Tell me, why i.e. the Xenix release before ~1987 (so before the product acquisition by SCO) point to only Microsoft, as per historic pictures here: web.archive.org/web/20130903075951/www.tenox.net:80/oses/xenix/ And even Doug Michels, Co-founder of THE SANTA CRUZ OPERATION, states like this in here th-cam.com/video/xZpaISQ82vA/w-d-xo.html (time index ~3:30): "[..] we have been working with Microsoft on Xenix [..]" "[..] it was easier to give Xenix from Microsoft [..]" "[..] we had a UNIX license from AT&T, but getting me the binary rights was very expensice was very expensive, and Microsoft had already done that. So we had a relationship with Microsoft [..]" This seems rather inline with what online sources, not just wikipedia, claiming on the history of Xenix and the involvement of SCO. It is undebatable, that SCO contributed to the code, and all sources point to that fact of joint-development by around 1982. There's also sources claiming about even two other companies being involved to work on Xenix for platform porting. [..] Pate, Steve D. (1996). Unix Internals: A Practical Approach. Addison Wesley Professional. p. 9. ISBN 978-0-201-87721-2. ""The Santa Cruz Operation (SCO) was formed in 1979 by Larry and Doug Michels as a technical management consulting business. [...] SCO then changed its focus from consulting to the custom porting of Unix system software and applications. The first version of Unix which SCO developed and sold was called Dynix, a name subsequently used by Sequent. The operating system was based on Seventh Edition Unix and ran on the PDP-11. [...] In 1982, a joint development and technology exchange agreement was reached between SCO and Microsoft bringing together engineers from SCO and Microsoft to further enhance the Xenix operating system which was increasing in popularity. Microsoft and SCO worked together with Logica in the UK and HCR in Canada, producing enhancements to Xenix and porting Xenix to other platforms." [..] For sure, SCO took it all over, making it their product by around 1987. But then action, to question it, would it be *their* (SCO's) product, if even two other companies (Logica, UK, and HCR, Canada, both later acquired by SCO in 1986 and 1990 respectively) were contributing!? But as far as I can tell from all the information available, Microsoft was acquiring the UNIX license in summer 1980, calling it Xenix, so Xenix was theirs in the first place. One might argue forth and back, if Xenix (with contributions by SCO), was an original product then, as one might argue about MS-DOS (QDOS), which was coming from SCP in the first place, and was essentially a CP/M clone. But then again, all I find is marketing docs by Microsoft, especially for the pre-1987 period. So if they did the marketing, then it was their product, independent of whom contributed to it, at least until that point, when Microsoft lost interest into it and sold it off to SCO. And should it be the other way around, then please, go ahead and update the online sources in Wikipedia and other retro computing websites with the correct facts and the proof to it. Until that is the case, I stick to the information available online, which is: XENIX was a MICROSOFT product (by 1980), which was later jointly developped between MICROSOFT and THE SANTA CRUZ OPERATION (from 1982) and even later sold-off to THE SANTA CRUZ OPERATION (by ~1987). And there is no single source claiming something like "XENIX was developed by SCO, but they let it market by MICROSOFT, just because ..." If anything, then SCO originally developped DYNIX. This is the way I see the situation.
At university, in the early 90s, I ported XBBS (a bulletin board system originally written for Xenix) to Linux. Good times!
Two further details - XENIX - like UNIX - was a virtual memory capable OS. This gave it more ability than a 640 Kilobyte limited MSDOS. Second - and unrelated - some viewers may not know but hard drives of the early 1980's (when the 5 1/4" form factor (same as floppy physical size) first came out did not have perfect magnetic surfaces. So drive failure was a common event and you had to format the full drive (usually 5 to 30 megabytes) and record the "bad spots" to avoid using them. Put that in perspective - a single PDF file with images can easily exceed that drive capacity.
It was VM capible, if the hardware supported it, which the 8088 version did not, it was a swap file situation, and the 286 version, that Unisys figured out how to put VM in DOS, MS still did not, it was not until the 386 versions that they supported VM.
The earlier 8086 version supported VM with special hardware.
Even though the drives I had ( a perfect ST-225, and a 2 error CMI ), I still turned off error correction and took 24 hours to test each drive ).
@@joeturner7959 not sure I would call their sandboxed protected mode a VM.
The FAT had an extra flag for defect sectors and turning on the VERIFY-feature (read cycle after writing, halving the usable speed of the disk) made sense, before IDE came out. On the other hand, the disk parameters were not fixed, but set in the BIOS, and people succeeded in over-formatting their e.g. 40MB MFM drive to 60MB, by using more dense cylinders and sectors, and speed up by testing the interleave factor.... these drives were more analogue tech, than digital.
@@martinb.770 Making denser cylinder and sectors increased the rate of bad sectors, which reduced the usable size, so it hardly ever was a perfect solution, and it also increased the corruption rate of the drive. You had to find a good balance, which usually was the manufacturer suggested values.
@@Bubu567 Sure, but it was possible and as HDDs were insanely expensive, some people tried and did test the limits.
Just the same happened at the end of the floppy era, when special tools pushed the 1440kB beyond 1720/1760kB, if compatibility to other drives was not needed.
Yup! That’s how we used tar for backups back in the day-direct to block devices like tape drives. This explains why a -f parameter is necessary to specify a file instead of a block device. Great stuff! Thanks for doing all of the work to bring us this walk down memory lane.
I worked at a Radio Shack store around 1993 or so, and the POS terminals out front were connected to a Xenix system in the back room that took care of all of that stuff. :)
TRS-80 model 16?
I just wanted to say that I have been really enjoying your videos, especially the ones about SGI! Keep up the amazing work ✨
Thanks! Glad you like it. Though no more SGIs for the time, I'm afraid. I'm currently in the mid-90s DOS/Windows period, and will soon do some stuff on networking. Hope you're going to like that as well ^^
@@THEPHINTAGECOLLECTOR Don't worry, as a fan of technology as a whole, I'm always interested in finding out about new things, so I'll be looking forward to those as well 😄
Back in the '80s I helped to administer an AmDOS (American Dental Office Systems) implementation which ran on Xenix. It was fun, and I learned a ton, since it required diligent care and feeding. Thank you for bringing back memories.
Oh wow blast from the past! Used to support Xenix and SCO UNIX at small businesses that had serial Wyse terminals and serial printers. Later discovered Slackware ( lots of floppy disks ) as first foray into Linux.
Same. We had Stallion serial multiplexors for serial for printers / dumb / "smart" terminals. Started with something called Yggdrasil Linux/GNU/X, or LGX, then Slackware etc. Then into Debian / Redhat.
Hehe you mentioned SCO, I followed the litigation soap in the naughties, Darl McBride was one slimy SOB. I've never seen a court case last so long without the accuser ever handing over concrete proof of plagiarism. They insisted linux had copy/pasted bits of code from SCO UNIX in it. In the end, SCO had nothing, they were just trying to cash in on Linux. Probably the most expensive chain of litigations ever, several subsidiaries pumped hundreds of millions into SCO, one of them being Microsoft - to this day I can't fathom why, but maybe you can shed more light on that?
Slackware was my first Linux distro.
Wyse 50s or Wyse 60s? The Wyse 50 could do a full 38400, whereas the Wyse 60 claimed it could do it, but the Wyse 50 beat it in races.
@@joeturner7959 Its been too long :) Looking at pictures though, they looked more like the Wyse 50 terminals with monochrome displays (I want to say green, but not 100% sure).
If I remember correctly, the TCP/IP package was a separate addon with a cost involved.
Only on 386, the 8086 an 286 used uucp over cio.
yes, and wouldnt let you use the same key on other systems on the same network
To this day I still have a SCO5 server running, Xenix was a little before my time but I supported a customer or two with them early in my career. A most enjoyable video...
I have a feeling that it is just delete and not ctrl-alt-del that will get you out of Snake, as del is the universal "get me the heck out of here" key on later versions of SCO Unix.
Never saw your channel until it was offered. I love vintage computer equipment. I have subscribed. Thank you for your hard work. Love Mr. Know-It-All
The software distributions for the Lisa are in floppy TAR format too!
It's surprising how fun is playing with those unix systems and how much of the legacy commands as "message" is still included and working on modern linux systems
I had Xenix on a Compaq computer for a project. The three of us could work faster on this one machine than each using our own. I loved it.
I too, argue with the Mr. Know-It-All in my brain. He usually wins, you're better at knowing things than me, clearly. Smashed that sub button, great channel sir!
Very interesting video! Thanks for uploading & _please_ show us more about Xenix on real hardware!
@Thiesi Sure, I'd love to see if I can get the dev tools installed somehow.
I worked with Xenix with some of our clients on Intel. It was surprisingly solid as an OS. Typically they were migrated to SCO UNIX. Maybe not so surprisingly as it was a licensed copy of AT&T Corp's UNIX by Microsoft. SCO OpenServer V was very stable and reliable, but lacked a lot of the latest Linux innovations like a decent shell, so you had to port it across or use Skunkware etc. Compared to Linux with some items, it was very slow with certain actions. We had some DEC Alpha servers, but they were on their way out at the time, extremely nice hardware though.
Xenix and other older Unix versions use UUCP with getty and a modem. They send around files via USENET with UUCP. I don't believe slip and PPP were even a thing back then. UUCP did everything.
Indeed, I saw the UUCP command while gazing through. I shall be definitely looking into this topic soon!
@@THEPHINTAGECOLLECTOR Indeed there was not TCP/IP networking on these versions (I used them at work). That came only later with the 386 version.
@@Rob2Yup! E-mail, Usenet groups and literally everything went via UUCP.
UUCP could distribute USENET-posts, and those posts often included files encoded with the UUCP-command UUENCODE. But USENET really had nothing to do with UUCP. UUCP doesn't need USENET. It's a kinda neat protocol, I use it daily. There's also IP implementations of it.
@@kanalnamn USENET started on UUCP (news batches), the NNTP protocol that runs on top of IP came later.
8 MiB in 1983 was... unfathomably huge outside of the biggest corporate and university IT departments! My first HDD was only 5 MiB and that was a mind boggling amount of space at the time.
By 91 to 93 I worked with Xenix. There was two servers (at the place I worked to). One was a 386-20, with 8MB RAM and 16(!) serial terminals hanging to it. The other one was a 486-66 with hopping 16GB RAM with more 16 serial terminals. They ran a extended version of DBaseII (Clipper like) called TSGBD. Never had an issue with the OS.
I guess you meant 16MB
@@ruben_balea That's right, 16MB.
@@bobalex6323 thank you for that that was highly confusing.
Interesting video, brings back memories of yesteryear 🙂........I'm waiting for the follow up video, this must be good without a doubt.....👍
I worked for a wholesaler that ran entirely on Xenix up until at least the early 2000's when I left there. Also had a fairly impressive old line printer that was insanely fast and would still have jobs backed up at times. Fortunately I was the sole nerd there and knew how to manipulate the print queue to make me the next job. Led to some confusion, I just played it off like I was dumb about that stuff as they were.
Also, With SCO Xenix and Unix, installing a dos partition on the Hard Drive, and installing MS-DOS on it, and then Xenix / Unix on the rest of the drive, allowed you to copy files between the Dos partition and Unix partition with commands in Xenix / Unix. In Fact, you can boot into either DOS or unix using the boot loaded that comes with Xenix / Unix.
I have seen these in action. You need a serial card in order to network it (the name of the card type still escapees my mind). Well, network in the guise of having serial dumb terminals, as I don't think it hadsny ethernet nor IP connectivity until much later.
Digiboards
@@SilvaD702
Those were the multiport 4 port and 8 port serial cards. I used a dual serial port card.
I've read about Xenix but I've never actually seen it till now. Wow.
I’m really curious about the other games, I’d love to see them all demoed.
Thanks for the video! Xenix8086 is a direct descendant of Unix v7, so no IP yet. Just serial ports and x-modem :)
@alextrusty2585 Thx, I was suspecting something like this already.
No, thee eras: Xenix 1 was Sys 6 based, Xenix 2 was Sys III based, and Xenix 3 was Unix 7, so
This was really System 7. Are you confused yet?
@@joeturner7959 As far as I can tell that's not correct...? I think it's like this...
1.0: Unix version 7 (no "System")
2.0: Unix version 7 (no "System")
3.0: Unix System III
5.0: Unix System V release 2 (SVR2)
@11:18, what is the "once" directory?
Xenix was used with Ungermann-Bass Ethernet cards (with twisted-pair adapters) and the XNS protocol (unswitched) for corporate LAN (along with IPX/SPX and netbeui). The xenix machines were used as mail servers with Microsoft Mail client software until exchange and outlook replaced them.
There was even an internal port of the xenix mail server to use TCP/IP on NT 3.1 so that the LAN could be segmented (exchange was delayed)
Oh, Ungermann-Bass, that seems to ring a bell. I‘m almost certain they were on stage in an early 80s episode of THE COMPUTER CHRONICLES.
I worked for Microsoft in the 90s, they did pretty much the same before NT was about.
Ctrl+Backspace can be used on later versions instead of Ctrl+C. PErhaps here too? You could try stty intr "^C" and see if that works for changing to ctrl+c
Good hint, I'll try CTRL+Backspace.
Used to use Xenix and SCO Unix back in the 80’s on a 386 with 3 or 4 dumb terminals hung on the back of it.
Any Z8000 Xenix systems around? That would have been my system of choice back in 1978, though the 68000 had more possibilities. In the end, around 1980, I got a Wicat 68000 with its own DOS-like OS. Chrislin 512K memory board which didn't work unless I blocked off the top 48K.
Portable terminal boxes. That brings back memories.
the TCP/IP tools are a layer on top of what ever networking you use to connect the machine. Its an era thing, T-base 10 (coaxial) was still a new thing with machines in the 90's, where slip was still very common, ptp (or ppp?) not so much
Truly wonderful. Thank you for the work to make this.
such a treasure trove.
Wow, my dad remembers xenix! Its funny, I use Linux day to day as my daily driver. the more things change, the more they stay the same.
Floppy problems aside, is there any chance to run a downloaded Xenix on "modern" hardware today? Mostly interested in having a working pre-K&R C compiler,
There was a version called Xenix 386, which propably works on later hardware up and until the early 2000s. Maybe not the present-days hardware, as it doesn't know anything about SATA and other modern days hardware. Would be worth a try.
Of course there is. Use a virtual machine like virtualbox or QEMU or a real emulator like Bochs or PCem. This is your only option on very new computers as they no longer have UEFI Legacy BIOS support. In addition, the A20 gate support on modern Intel mainboard chipsets is missing too, but that shouldn't be that important for Xenix.
slip and a 1200bd modem is where i started … .. felt kinda magic at the time. Tar was The most incredible swiss-army knife at the time… especially when I later got some Sun workstation with a tape drive .. … too bad I gave those away… but it was fun to run linux/x11 on them in the 2000s …
Back to the 80's, the best operating system from an architectural point of view was QNX 2.x. Multi task, multi user and realtime ! There was no Ethernet networking (TCP/IP), but setting up an Arcnet network was a piece of cake and you could easy access all the ressources from a remote machine as you would for your local machine. An OS ahead of its time !!
Remember seeing the Xenix disks offered with the Radio Shack TRS-80s. Never used it.
What an awesome channel - how did I miss this? Why do you only have 3,5k subscribers?
i used to use SCO XENIX in the late 80's early 90's... if you have a dual operating system you can type dos at the xenix boot prompt and it should boot into the dos partition without having to swap active partitions with fdisk.
i have worked with Xenix on a trs80 model 16 back in the days.
That would be Xenix era 2.
You’ll need a 486 with the zorp module to upload to the Ethernet from Xenix.
cool stuff. i remember installing minix to my hp200lx using a pcmcia disk back in 1995.
Nice trip down Memory Lane.
I did sysadmin work on a Xenix system in the mid 90s
You‘re welcome!
i cut my teeth on Xenix way back in the 386 days. they were used to run major call centers alongside serial terminals, message boards and sat hanging off the PBX switch. I still have the SCO cards with the license numbers on them and still have a couple of VMs running Xenix, Unixware and openserver with their respective dev kits
How could you resist playing Rogue?
@anthonygelona9479
You would make a good MR KNOW-IT-ALL as well ;-)
Honestly? I had to make a break to get the video out. There's so much stuff in there to explore, too much after all! :-)
It came with Rogue, but BSD rogue 3.6 - hjkl keys.
I've played around with Xenix on the Apple Lisa some. The micnet networking was interesting a friend and I got that set up between a Lisa and an IBM XT.
XENIX was my first Unix experience. I know it was influential and Microsoft did not get enough credit for its pioneering. When SCO Unix acquired it I wondered if their marketing should have included "We sell C shells by the seashore." ;)
We had about 400 Xenix servers running along with about 300 Netware boxes. They were very reliable boxes.... Would reliably crash daily.
The entire air-services radars in Australia used Xenix - and was quite good for the time. The Netware products were significantly better for file and print. We ended up migrating to Linux over a few months. Much more reliable.
Wasn't there a micro emacs for Xenis?
Used to store raw tar files on floppies instead of filesystems, in addition to splitting them with dd. It's was so much faster & more efficient for large files.
Great job! It would be nice to see some IBM PS/2 back to life... could you? :)
@Pinguinomannaro74 I'm not so much into desktop systems, for the most I collect portable systems. There were though IBM portables from the PS/2 era, notably the P70 and P75. I'm looking around to get hold of one of these since a long time. And the day I get one, there will be also a video about it, promised.
Just wondering if it has multi console login buy alt-f key coherent unix and sco both support this in the 80s
Yes, but the multi console came at the cost of 3.6k ram per console. ( the manual said 4k ) so I had only 2 on my PC, and later 8 on a 386...20
I did my intro to C course in a class of 20 terms all running off a 386SX40 2MB and Xenix
Super video.
At these times it was not uncommon to write tar, cpio or dd directly to the external media.
If I remember correctly Interactive Linux was delivered that way ....
Amazing to watch in some aspects. One side of the Unix medal screams about complexity of Unix for weak 8086-not-so-weak-286, another point why MS ditched Unix, they felt like creating something better and simpler. Dos was fast, small and glory to that marketing department who envisioned easy tools for programmers and lots of software for users.
DOS and Xenix were meant to compliment each others, they existed side by side and catered to different needs. It was co-operation with IBM on OS/2 as well as Windows (which originally were meant to coexist as well - with very similar idea). DOS and Xenix were the old plan - OS/2 and Windows were the new plan meant to replace them.
And really, DOS lasted way longer than it had any justification for doing so ;)
With SCO Unix and its tcp/ip package, that package needs a key as well and you couldnt reuse the same key on systems that were on the same network ....
If you are not supporting large binaries, you can shrink the swap area to 1280, ( 640k/512b block sizes). If running large binaries, try a larger hard disk ( My hard disks were 33Mb, and 31Mb, with a 2560 swap size on its own partition. It did not help ).
( you also forgot to empty the /temp folder of the tar files.
For TCP/IP you need a 386. For GNU Tools, and a decent compiler, NOT The MS 'C' Garbage that MS got from Lifeboat., you need a 386.
The disk images are up on pcjs for the dev tools, and I used them to get through my 'C' class, and pass, but...
The teacher started with The Lifeboat compiler, and a student complained about something...
the teacher sat down at the students computer, and typed a simple program,
and compiled it. He looked at it again, and then
picked up the box for lifeboat C, and threw it across the room at the garbage can,
and announced that we will be using another compiler, next week.
No worries, I did clean the /tmp, it was just not shown in the video as not adding value
It's most certainly "networkable" - I recall seeing a mention and possibly even screenshot of Internet Explorer on Xenix; Don't ask me which version, must have been really early... I've always been interested on whether the Xenix IE port was 16-bit and able to run on 286 CPU's, or if it required the 386 version?
What I was most interested about, and was bummed out because the video didn't cover any of it, was the software implementation for things like multitasking, swapping, etc. - like did it use 286's ability to do pre-emptive multitasking or did it rely on co-operative multitasking; and what about on pre-286 CPU's, was there any way around having to use co-operative multitasking and let the application steal the control by simply running as long as it wants without returning the control back to the OS?
And what about swap space - I notice that it asks you for it, but on a hardware without virtual memory support that would allow swapping to disk, so is that swap space going to be just waste of disk space, or did they implement some crazy scheme of "co-operative memory management", where the application can tell the OS that "this memory can be swapped out" and then everytime it wants to use it, it has to call the OS to make sure that the stuff it wants is in fact in RAM? I mean, that's just a crazy example of what I was thinking as solution for just-for-fun OS project that would run on ancient PC's, I just mentioned it as example, but whatever Xenix actually did or didn't is what I'm interested about :P
I'm assuming, partly supported by comments I already glanced at, that the swap space wasn't useful on pre-386 hardware, but then why would this version of Xenix ask for it? Or does it include support for 386 and operate in protected mode if available - and if so, I don't think the user space applications included had two separate versions, so then would it run them in Virtual x86 mode and the applications be just 16-bit real mode code?
Oh man, I could just go on asking questions and speculating - it's a really bad habit of mine, this sort of questions are better asked one at time, so you can get an answer and don't have to ask questions based on erroneous guesses about that first question... Like if I just waited to find out if it even used the swap space on pre-386 CPU's, then I wouldn't have to ask speculative questions of how did they implement support for swap on 286's or even earlier if the answer was that they didn't :D
Still, thank you - this is the only video I've seen about Xenix specifically, I've only seen it mentioned briefly as UNIX system for early x86's without it even being shown in use (and not just some old marketing screenshots).
During the 80ies / around DOS 3.3, it was still speculated, that MS will enhance DOS to being Unix-compliant by DOS 5.0.
Needless to say, it never happened, and even POSIX-compliance took until NT 3.51, in the mid-90ies.
although you might have to construct the file lists by hand, you should be able to down grade the 1.2MB install disks to 360KB disks, but you will need the "tar disk format" tool to do it properly - that should be present somewhere, but may also only be on the Developer Tools ... before doing that, I suggest temporarily hooking up a 1.2MB drive, and seeing if your disk images are good, before wasting time on a dead end (due to bad images) - the "tar disk format" tool may already be installed with the basic (extra) Operating System install
Good hint, will definitely look into that! :)
Os melhores sistema operacionais do mundo até agora!!!
Xenix was technical wonder back then. Single intel 286 PC can serve 8-16 concurrent terminal or telnet session user comfortably. It was good and cheap departmental office machine. We developed ditroff backend for Canon laser printer (300dpi) to print beautifully typeset document for research paper and books in 1986. Way better than CP/M WordStar word processor with Diablo daisywheel impact printer.
It had rogue... Exactly what you need
Yes :) I used to have tar on floppies, too even with -M :)
This is fascinating stuff. A bit above my level, but informative nonetheless.
Microsoft didn't sell MS-DOS on its own, until version 4.01. Until then, Microsoft licensed it to vendors. The same situation with XENIX, and OS/2.
I think the youtube algorithm likes this video, there were 645 subscribers and 4k views when I started and 717 subs and 6.1k views when I finished. I also used Xenix and SCO3 and 5 in the 90's.
Yeah, I’m totally blown away by the attention this video receives. In relation to my small niche channel, views and subscribers are totally skyrocketing! Seems like everyone is into Xenix :)
not sure if this was said already, but it bears mentioning that xenix was a real-time OS, unlike most 'nix's of the time. we used it for control systems and such that demanded predictable responses, when hardware was slow and the OS could get bogged down by a single poorly written app
omg you didn't even show what would be the most interesting aspects of running Unix on 16bit chip without any protected memory! did they implement this in software somehow? Is all disk I/O done through the PC BIOS? Show us the weird!
They implemented VM and swapping, no protected memory. The c compiler destoryed my file system 3 times, and I gave up.
The graphic with AT&T, Unix, SCO history is very good but confusing.
Best wishes.
Good, but confusing. :-) The entire history of it is confusing, if we think about it. The thing could be even more confusing, considering
(a) Novell actually selling of the UnixWare product, coming out of the UNIVELL joint-venture to said Caldera, later rebranded THE SCO GROUP (though essentially keeping the UNIX copyright)
(b) The entire DRI (Digital Research) timeline, including CP/M and DR-DOS, eventually transitioning via Novell over to Caldera, Lineo, DeviceLogics and DR DOS Inc
It's all a big huge labyrinth of acquisitions, mergers, splits, rebrandings and more over the past decades.
@@THEPHINTAGECOLLECTOR I think it can be improved / explained in a clear way.
Best wishes.
I'd argue that compilers still isn't included in (fairly) modern UNIX OSes. AIX which we still use a lot at my company requires a license. Solaris used to be the same the last time I used it in a serious fashion back around 2005.
(You obviously had the option to use GCC but that required other libraries etc to make the executable portable between machines)
I was a bit unprecise in my wording about "included". I always ever had access to the full disk sets, which included the compilers to install.
It's the same on SGI IRIX as well, it came with the development foundation and compiler CDs, and they could be installed right away.
It's correct though, you still needed the extra license to use it.
Ha, you would have hated AIX 2.x then - never mind the development tools, the *man pages* were a separate layered product…
/usr is not slas user but
unix system resource
Please see the statement concerning that topic further below in the comments.
I have never, in my life, heard the phrase "throw a look", and I've been processing it for way longer than I should be.
This way of saying comes from the German language and should actually be translated into English as "Take a look at Xenix".
@@OpenGL4ever I assume you mean Schauen Sie sich, but I have no idea what Xenix is, I also don't know how you get "throw a look" from that either.
@@Melechtna Such a word "Schusauen" doesn't exist in German.
The English sentence "Take a look at sth." is translated to "Wirf einen Blick auf etwas." in German.
a = einen
look = Blick
at = auf
sth. = etwas.
But the difference between "take" and "wirf" is, that "take" means "nimm" in German, while "wirf" means "throw" in English.
Or in other words, no one says in German: "Nimm einen Blick auf etwas."
Xenix is a type of Unix operating system for the x86 processor and was used in the 80s before Windows NT.
@@OpenGL4ever That would be because my phone tried correcting it several times, and it didn't like me typing Schauen, and I failed to correct IT the 6th time.
@@OpenGL4ever Also, why are we using Xenix as an example of an object to be looked at, that just seems like a really bizarre and random thing to throw in there.
I remember Xenix and x.25 packet routing in highschool
1990's 386 with Xenix, 4 x Wyse 50's. Wordperfect. A multiuser WP center.
For the failed mfm drive you could use one of several mfm emulators, not cheap
I would have used an XT-IDE adapter, less expensive ;)
Though luckily not needed as the drive could be revived and so the machine remains fully in original condition.
XENIX developers where maybe the only ones who managed to use the 80286 crippled memory management and protected mode implementations more or less ...
Everybody else went away in horror and came back when the 80386 was released.
What about OS/2? I'm not absolutely certain if it was DOS applications that it was used for (that's what I recall), but it ran in 286 protected mode, and unlike Windows (which would just switch to fullscreen text-mode and give the control over to the DOS application and resumed when it finished - plus I don't think it ran in protected mode; I'm talking about 3.0 that we had on our 286, though if you had a 386 you could run it in "386 mode" and multitask between Windows and DOS applications), it could even multitask between real-mode DOS programs and OS/2 programs - of course Xenix didn't have to switch between real-mode and protected mode, but I'm mentioning that feature to emphasize that it ran on protected mode. If not obvious, I'm talking about 16-bit OS/2 that ran on 286.
Also, I'm not 100% certain, but almost anyway, that Minix (either 1.0, 1.3 or 1.5 I believe) used 286's protected mode.
There were others as well, including some less known DOS applications that enabled multi-tasking, or at least task switching, on 286 and I believe some competing DOS operating systems.
But I'm moving on increasingly uncertain areas, so let's just say that OS/2 is definite proof that Xenix wasn't the only one.
@@robsku1 Windows 3.0 ran in protected mode on a 286. In Windows 3.0 this was also known as standard mode. Windows 3.0 also had two more modes: Real mode and Enhanced mode. Real mode runs on an 8086, enhanced mode requires a 386 or better.
Enhanced mode took advantage of the 386 CPU's VM 8086 mode for DOS applications. Thus, you could run multiple DOS applications in their own VM. The VM instances ran in preemptive multitasking, the Win16 applications only in cooperative multitasking. I'm not sure about Win32s applications.
In Windows 3.1 there was no more real mode support.
And in standard mode you could only run one DOS application at the same time. To do this, Windows 3.x switched back to Real Mode and DOS. It wasn't just a fullscreen text-mode, it was DOS using DPMI, while Windows 3.x was kept mostly in extended memory.
OS/2 did also run in protected mode on a 286 and was also limited to only one DOS application at the same time. The 286 lacked the VM 8086 support, which was required for more.
Windows NT was Microsoft's first OS designed purely for the 32-bit protected mode of the 386 and later.
Great✌✌
My business name was Zenix Consulting in the 1990.
I dealt with Xenix once... On an ANCIENT terminal setup in a pharmacy. I might have cried. It was awful. And this was in the late 90s...
I remember the Rodime drives being not very good back in the day.
Blast from the past...I have only seen one Xenix system decades ago. I think it was a Unisys with a bad MFM drive....if my memory serves right
You can get more out of your hard drive if you use a RLL (run length limited) hard drive controller as old hard drives are basically just dumb devices you can cram 1.5 times more on the drive. I am not sure that is helpful. But I try to be helpful.
remember it very well. 286 1mb ram. sync;sync;haltsys
Seems the Internet was a little bit too late for Xenix. Berkeley Sockets wasn't developed on BSD until 4.4BSD, in 1983. Cu should be available though, with uucp. The wonders of slip and later things, you may have to compile.Unix of the 1980s was clearly a far cry from even lowly Linux, lol. How much difference 13 years makes.
At SCO once I was visiting and some dude walked in from the hot tub completely naked, walked to his chair, again naked, and started coding 🤠👍
That qualifies as things you desperately want to "unsee" ... ;)
@@THEPHINTAGECOLLECTOR nah I'm a humanist and it was sort of funny 🤣
Let's keep it forgotten...
WTF is with Microsoft putting X in everything?!
Look, if Xenix hadn't suck back in thevday then Linux would have never been bourne 🤪 (the pun was intended)
A bit harshly said, but true. Unix at the time was too demanding, and didn‘t establish well outside specialized niches and the server market. And licenses were to expensive for most private users, which indeed was one of the arguments for Linus to create the Linux kernel after all. Now, Linux is everywhere.
But let’s not forget that there is still a descendant of the original AT&T Unix out there in the consumer mass market, which is Apple‘s macOS with all its iOS derivates, accounting for 1.8 billion devices out there. So while the Microsoft Xenix branch ultimately was a deadend, the BSD-derived macOS (NeXTstep) branch has made it.
Its sco.. though a failure.. it was a popular product up to about 10 years ago.
xinix > sco > unixware > now linux? 😂
Nice video, but this constant interruption by your alter ego, and especially the windows closing sound at the end of the interaction, is a bit annoying.
Please see the community tab and the poll related to that topic. Open to suggestions.
Talking to yourself is a manifestation of a split mind.
Wanting so badly that Microsoft ever has made an operating system all by themselves is impossible. Everything sold by Microsoft was made by other persons or companies. It was bought for a few pennies from the silly bastard that was not able to bring the product to the market. Unix was made by the telephone company, BASIC was made by some school. DOS was a copy of CP/M, Windows was found somewhere in Palo Alto at some research company.
"Unix was made by the telephone company": correct, by AT&T and the Bell Labs (licensed)
"BASIC was made by some school": Correct, by some people at Darthmouth College (didn't find if that was licensed, bought, or whatever)
"DOS was a copy of CP/M": Correct, by Seattle Computer Products (licensed, later acquired)
"Windows was found somewhere in Palo Alto at some research company": Nyah, not sure here if they actually took (as in bought, licensed) code base. For sure, strongly inspired by the XEROX ALTO, found at XEROX PARC, yes. Apple did the same, btw.
"Talking to yourself is a manifestation of a split mind. " - Well, no. I'm talking to an avater inspired by Microsoft Office 95's clippy, which is just totally annoying.
But I seem to be getting the point about what the other guy commenting yesterday meant about it, that it disturbs people ... Well ...
I guess NT was bought from DEC, then? Doesn't matter if Dave Cutler was paid directly by Microsoft to lead a team and code an OS from scratch...
@@TiagoJoaoSilva It differs a lot! It is the principle. If an idea for software is developed with good standards by people with a right mind, then the result could be something that is good and useful.
If an idea is bought and then further developed by others, then that company's goal is to commercialize ideas of others. That is a totally different starting point then problem solving with the help of computers.
Because there is a totally different goal, the result of both kinds of software is not the same. All software made by Microsoft needs well trained technicians to install and operate. Software for the Mac does not need technicians to work on it, the aim was to make software for creative people that can be used intuitiously.
Nice. Don't say user when talking about /usr. usr stands for Unix System Resources not user.
@galier2
/usr in my understanding goes OK for "user", as that's what the Linux documentation project also calls it: "User System Resources"
One may correctly argue that Linux is not old enough to claim naming it, but also the BSD-line hier(7) man page [man.freebsd.org/cgi/man.cgi?hier(7)] always ever refered to /usr as "contains the majority of user utilities and applications", and the BSD traditionally contained even the home directories beneath /usr/home.
The UNIX Version 7 hier(7) as the oldest I could find, refers to /usr simply as "general-pupose directory, usually a mounted file system" [www.tuhs.org/cgi-bin/utree.pl?file=V7/usr/man/man7/hier.7] with no reference to it being "Unix System Resources".
I stumbled accross this one here [unix.stackexchange.com/questions/103347/pronunciation-for-usr-directory], I found this claim:
"In this 1982 film from AT&T that features the UNIX inventors, The UNIX Operating System, Brian Kernighan pronounces /usr as "user", around the 13:41 mark"
Now, honestly, I don't know. It may well be that it originally was intended to be "Unix System Resources" some 60 years ago, though the only notation I personally ever had been confronted with in 25 years, was the notion above about the directory containing user-related data and binaries.
According to my not very indepth research, there is some controversy around it.
/usr was originally used, as I understand, for stuff not needed until multi-user mode was brought up.
I think "Unix System Resources" is a backronym.
@@fnjesusfreak Indeed.
Originally the user home directories were under /usr as well... /home was invented later.
@@THEPHINTAGECOLLECTOR Ok. I stand corrected.
I've heard /usr been referred to as "usser" (not "user") from several poeple. I'm not sure if they have invented that by themselves, or if it's common elsewhere.
#721
it's not really 'long forgotten'. it's basically multiplayer dos or cp/m and as such most likely parts of it made it into the os/2 kernel and are now simply called 'windows 11'. with some fancy new gui and new colors. it's just reselling the same shit over and over again under different names. it is also not really quite a 'unix unix'... more in the league of other programs that have dos/cpm api calls and can run multitasking, such as that quarterdeck stuff. xenix has barely any networking capability for one thing. but hey. os2 which later became windows nt 3 had to come from somewhere no? LOL. this is not something that is in the leauge of things like sys-v or dec vax. this is more like a glorified task swapper.
I strongly disagree here.
The early releases of Xenix were definitely a direct evolution from licensed AT&T UNIX Version 7 source code.
Still, it has nothing to do with either of the DOS, CP/M, and later OS/2 or Windows NT line of operating systems.
Attributing Xenix of beeing a "dos/cp [..] task switcher [..] like quarterdeck" is inadequate.
It is correct, that DOS was essentially a clone of CP/M, of which naturally a lot was inherited into the DOS VDM present in both OS/2 and Windows NT.
However, the OS/2 kernel clearly is neither DOS, nor CP/M, and not of UNIX origin either.
Windows NT has indeed taken some inspiration from OS/2 and even more from VMS, but again is neither a DOS, CP/M nor UNIX descendant.
And if you watched closely, the SCO Xenix release 2.1.3 showcased in this video reports itself as UNIX System V, which originally came out in 1983.
So Microsoft and SCO definitely had updated their source base from UNIX Version 7 to to System V at some point.
It is also well-known that Xenix System III as being an intermediate release, was derived from AT&T UNIX System III.
I didn't find the earliest Xenix release to play with, but having one would definitely exhibit an even more archaic look and feel than the System V release we've seen here.
Actually, it says right out it's System V.
@@fnjesusfreak then why is it's entire range of int21 functions essentially the same as cp/m and dos... :P i'm pretty sure no actual unix ever had that :P
ofcourse one can damn well make that reside on top of the task scheduler borrowed from any actual unix... but ya know. it's still just multiplayer dos :P
@@CB3ROB-CyberBunker Can't be the same as both, as neither is the same as each other.
XENIX IS NOT, REPEAT *_NOT_*
A Microsoft PRIDUCT!!!!!
Get your facts straight, kid!
„Xenix is a discontinued version of the Unix operating system for various microcomputer platforms, licensed by Microsoft from AT&T Corporation in the late 1970s. The Santa Cruz Operation (SCO) later acquired exclusive rights to the software [..]“
Source: en.wikipedia.org/wiki/Xenix
It is explicitely stated in the video, that Microsoft licensed UNIX from AT&T.
It is correct, that Microsoft didn‘t invent UNIX, and no such claim is made!
But still, Xenix was Microsoft‘s Unix version for the time, though they couldn‘t use the UNIX name, as they only licensed the source code, hence they named it Xenix.
If anything, then maybe you should check the facts as well.
@@THEPHINTAGECOLLECTOR exactly...
SCO developed XENIX.
Despite what is in wikipedia....
@@Theineluctable_SOME_CANT i wasn‘t there, but all sources claim otherwise…
en.wikipedia.org/wiki/Santa_Cruz_Operation#Relations_with_Microsoft
„Xenix is a Seventh Edition Unix-based version of the operating system that Microsoft worked on, initially for the PDP-11.[3] SCO first began working with Xenix in 1981.[8] In 1982, Microsoft and SCO forged a joint agreement for development and technology exchange, with the two companies' engineers working together on improvements to Xenix.“
„In any case, intellectual property rights were not transferred in the 1989 agreement and SCO would continue to pay Microsoft royalties for Xenix and Unix technologies.“
„In 1983, SCO made a technically difficult port of Xenix to the unmapped Intel 8086 processor (earlier 8086 Xenix ports required an off-chip MMU) and licensed rights from Microsoft to be able to ship its packaged Unix system, Xenix, for the IBM PC XT.“
To me, this does not read like „xenix was SCO‘s“ but more like „it was Microsoft‘s“.
For sure SCO contributed to the code base, and for sure, they later acquired Xenix from Microsoft.
But wether SCO was there from the beginning on or not, I find no claim it was their original product.
Also, the particular version of Xenix System V v2.1.3 shown in this video shows the Microsoft copyright clearly predating SCO‘s copyrights.
@@THEPHINTAGECOLLECTOR I was there. I installed XENIX on a number of systems...
@user-sf5bz5nh9e
"Being there" implying being a SCO employee, so having first hands hard fact information?
But then please, proof it. Tell me, why all sources tell different.
Tell me, why i.e. the Xenix release before ~1987 (so before the product acquisition by SCO) point to
only Microsoft, as per historic pictures here: web.archive.org/web/20130903075951/www.tenox.net:80/oses/xenix/
And even Doug Michels, Co-founder of THE SANTA CRUZ OPERATION, states like this in here th-cam.com/video/xZpaISQ82vA/w-d-xo.html (time index ~3:30):
"[..] we have been working with Microsoft on Xenix [..]"
"[..] it was easier to give Xenix from Microsoft [..]"
"[..] we had a UNIX license from AT&T, but getting me the binary rights was very expensice was very expensive, and Microsoft had already done that. So we had a relationship with Microsoft [..]"
This seems rather inline with what online sources, not just wikipedia, claiming on the history of Xenix and the involvement of SCO.
It is undebatable, that SCO contributed to the code, and all sources point to that fact of joint-development by around 1982.
There's also sources claiming about even two other companies being involved to work on Xenix for platform porting.
[..] Pate, Steve D. (1996). Unix Internals: A Practical Approach. Addison Wesley Professional. p. 9. ISBN 978-0-201-87721-2. ""The Santa Cruz Operation (SCO) was formed in 1979 by Larry and Doug Michels as a technical management consulting business. [...] SCO then changed its focus from consulting to the custom porting of Unix system software and applications. The first version of Unix which SCO developed and sold was called Dynix, a name subsequently used by Sequent. The operating system was based on Seventh Edition Unix and ran on the PDP-11. [...] In 1982, a joint development and technology exchange agreement was reached between SCO and Microsoft bringing together engineers from SCO and Microsoft to further enhance the Xenix operating system which was increasing in popularity. Microsoft and SCO worked together with Logica in the UK and HCR in Canada, producing enhancements to Xenix and porting Xenix to other platforms." [..]
For sure, SCO took it all over, making it their product by around 1987.
But then action, to question it, would it be *their* (SCO's) product, if even two other companies (Logica, UK, and HCR, Canada, both later acquired by SCO in 1986 and 1990 respectively) were contributing!?
But as far as I can tell from all the information available, Microsoft was acquiring the UNIX license in summer 1980, calling it Xenix, so Xenix was theirs in the first place.
One might argue forth and back, if Xenix (with contributions by SCO), was an original product then,
as one might argue about MS-DOS (QDOS), which was coming from SCP in the first place, and was essentially a CP/M clone.
But then again, all I find is marketing docs by Microsoft, especially for the pre-1987 period.
So if they did the marketing, then it was their product, independent of whom contributed to it,
at least until that point, when Microsoft lost interest into it and sold it off to SCO.
And should it be the other way around, then please, go ahead and update the online sources in Wikipedia and other retro computing websites
with the correct facts and the proof to it.
Until that is the case, I stick to the information available online, which is:
XENIX was a MICROSOFT product (by 1980), which was later jointly developped between MICROSOFT and THE SANTA CRUZ OPERATION (from 1982)
and even later sold-off to THE SANTA CRUZ OPERATION (by ~1987).
And there is no single source claiming something like "XENIX was developed by SCO, but they let it market by MICROSOFT, just because ..."
If anything, then SCO originally developped DYNIX.
This is the way I see the situation.