In September 1968 I was sat at a Teletype terminal creating print and 8-hole tape copies of an Algol program for an Eliot something computer in the room next door. A year later I was using a 80 column card device to produce Fortran code to run on a Univac batch processing machine that was a 45 minute train ride away from the office. No mention of operating systems yet. My first OS encounter was on a Ferranti computer controlling a nuclear power station - an interrupt driven system using a physically huge (wardrobe size cabinet) drum for secondary memory. That was the foundation for working with desktop Personal Computers from 1977 and following all these names like CP/M, UCSD p-system, PCOS, MS-DOS, Windows, and UNIX (several flavours). This video seemed to map my leaning curve over almost 60 years in computing - as @KangJangkrik wrote 3 hours ago an engineer's bedtime story. Thank you for this rewind.
I can relate. I started programming as a high-school sophomore in 1976. After passing the first course writing programs in BASIC on a Model 33 Teletype with paper tape storage I was the only student that year to learn FORTRAN. That involved going to the school district administrative building to use their card punch, then taking my card deck to the data center, and finally picking up the results the next day.
@@kyriosity-at-github Re 'shuffling'... ...and that is why many languages and their compilers in the punched-card era had optional line numbers to allow manual sorting. That said, having access to either a card-punch like the glorious IBM 029 which could print the contents on the top edge, or a dedicated 'interpreter' machine was a luxury to us progammers who usually "made do" by drawing a diagonal line across the top of the card deck with a marker pen. No - we didn't use sealing wax.
Just think about this for a moment... all this has happened in less than a hundred years, from the first telegraph to having pocket computers and phones in the palm of our hand, which interface with AI/LLM's and who knows what more to come.
Never had I thought I'd hear about the human centipede on this channel, let alone using that as an analogy to Unix pipes. You've outdone yourself there, Jon.
I suspect that analogy has some staying power, since it does render the idea both rather accurately as well as in a funny, easy understand and visualize way.
The story of the use of computers in general is layers of abstraction. Abstraction on top of abstraction on top of abstraction, each layer allowing faster development, though also adding a layer of overhead. There's a *lot* of overhead. Finding ways to reduce that overhead will help mitigate the end of Moore's law.
@@brodriguez11000 I have a few minor contributions in the Linux kernel. Hardware is outright hostile, and debugging is frustration as an olympic disciplin. Software, while as error prone as math, has incredibly powerful development and debugging tools. The problem just is that if something is comparatively easy, humans push the envelope until it becomes hard.
10 print "obscene word" 20 goto 10 run My first program for BASIC when I was 7. I was very impressed the first time I saw windows and a mouse as a kid. I tried to explain it to dad at supper; he didn't pay any attention.
@gafakyusef6201 No, Industrial Electrician, Automation, Refrigeration, and HVAC. Programming a PLC for automation is similar to basic; especially if it is a Beckhoff twin cat and written in script. I prefer ladder logic and Rockwell Automation with Allen-Bradley plc; ladder reminds me of basic. I wrote many HMI programs to tie to my PLC programs; I think learning basic as a kid made it easier for me than for the older generation, when automation became more common.
"Its like the human centipede of computer processes" "... written in the high-level C language" "So there was Kildal, in his room, with just a naked floppy drive" Goddamn our boy comin in HOT
Not sure what your issue is with "... written in the high-level C language", given that at the time most of the stuff under the hood was written in assembly. While "C" is considered pretty low level now, it was not at the time.
@@thekinginyellow1744 As you say, it's probably the historical irony of how abstract and high-level C was viewed at the time, compared to its current view as being merely a step above Assembler.
@@watchm4ker I wouldn't be surprised if all of today's high level language are gonna be seen as low level in a few years, replaced by programming through AI
Retired electronics eng here. There was a story that at one point when MS was trying to kill Lotus 1-2-3 that the OS team had a saying for the next version of DOS - "DOS aint done til Lotus wont run". But that was by far not the worst thing that MS did. Often they would come to companies that had software like a TCPIP stack (I worked for a company with the first commercial stack for the PC) and say we want to license it for 20 cents a copy (we got $200). They would say either accept it or we develop it and include it in the OS. This was how much of windows utilities originated - disk utilities, FAX software, TCPIP stack etc. One by one they got put under by the juggernaut of MS. It meant more for less for consumers but often it also meant an inferior product. I can remember a large airline begging with us to not stop supporting our windows TCPIP stack because the MS stack just didnt work in their environment. BTW we also had the first browser for a PC called Emissary and its logo was a blue e - sound familiar?
You should consider writing a book. There are so many out there about the broad strokes of the PC history, I’d think there’d be a enthusiastic interest in one ff stories like yours, told from the perspective of someone who lived it.
You have taught me more IT stuff / history than I learned in my first year of Electrical Engineering. Thank you. I wish my Samsung phone had my old Palm's Graffiti.
I think Raymond said it best. It was trying to be the perfect solution to a problem no one had. Unix already existed, and had moved well past those ideas.
@@YanestraAgainbelieve it or not plan 9 is used for wsl for its protocols involving invoking Linux and windows integrations with the file system and io
I was Superman when writing a big, functional MS-DOS batch file in early 90's. I know this may be silly, but to me personally, that was the joy of computing. Thanks for this amazing video, mister.
The amount of knowledge this channel provides for free is insane. I'm definitely subscribing next month for patreon. Thank you for hard amount of research and effort into these videos
Well done John in making the distinction between operating systems and IPLs, kernel-mode programs, and HALs. Many operating systems, antique and modern, don't fit directly into any of those boxes. Well done also for picking points that most people will connect with while still keeping it to 30 minutes, too. This could easily be a 20 part series if you wanted it to. I particularly like that you've addressed the economic impact of the "IBM PC" marketing. The rise of the PC in the face of cheaper and more powerful options has always puzzled me.
I think Tanenbaums definition is the best (he probably didn't come up with it, but it's what he states in one of his books.) He says that an operating system perfoms two functions. One is resource management and the other is to provide an abstraction layer. A sort of extended machine. If you use the definition most people use (erroneously in my opinion), then you'll end up having to argue that Edge is part of the Windows OS. Which is pretty ridiculous. It's just a program shipped with the OS.
It reminds me of an exercise problem in the Silberschatz / Operating System Concept book. (Paraphrasing more or less) Who decides which is and is not part of an operating system: the user, the experts, or the court system?
I argue that the browser has become its own OS, as are cloud-provider-level microservice aggregates. The only limit to the OS turtle-stacking is theoretical.
@@JohnnieWalkerGreenI think the experts. Most users are quite uninformed and probably most judges too. The idea that Linux or Windows is just everything that comes with an install (in terms of calling them OSs) is very nebulous and reductive. I think most people use such a definition because they don't know any better and then if they ever have it pointed out to them that there is a stricter definition (I'd also say arguably more correct and useful) they don't want to accept or consider it because they didn't come across the idea on their own. Obviously that's just my opinion, but I do think it's not a very good definition. I mean if they remove Word pad from Windows, but keep everything else the same is it then a whole different version of the operating system? I wouldn't say so at all. Word Pad is just a user space program that shipps (or shipped) with Windows. Also the definition I gave is an actual definition. What would you say is the definition of an OS that most people use. Yes you know it when you see it (sort of), but can you really describe it succinctly and clearly. Where as if it's just some low level software that provides two distinct but useful functions that basically any modern system that people would refer to as an OS provides at it's core, then it's relatively easy to define. Anyway I know that was a bit of a rant 😅.
@@poofygoofI don't think so. It's similar to when people say Emacs is like an OS. It's just an interpreter. Browser are similar to OSs in some ways, but I wouldn't go as far as to say that they ARE OSs.
@@hamesparde9888 the distinction is arbitrary -- what makes a LISP machine from the 80s an OS but EMACS LISP not? DOS and CP/M didn't have much in the way of resource management, but don't they count as OSes?
I'm sorry that the Commodore AmigaOS gets left out of these conversations. It's preemptive multitasking would have fit in to this video well. Excellent video, as usual!
@@Longlius sacrilege! I did more in 8mb than any mac or pc of the same era... add arexx for interoperable software hooks, and it was like having a superpower across my little renderfarm. I even ran photoshop in emulation faster and more efficiently than the expensive mac quadra with insane ram. For a time, it was the absolute best for multiple pro use cases.
@@briancase6180 That's valid, it was based on unix. And the Amiga's that shipped with unix was using a port of AT&T system V. Sun microsystems and Unix international used to show on Amiga's at trade shows.
Symbian OS was born from the PDA world, focused on optimising limited hardware resources. Programming on it had a very steep learning curve, due to this optimization and the absolute difference from typical PC programming. It was very hard for any programmer not experienced with Symbian to move to it and port any of the existing software. No matter what we did to improve the tools, it was hard to program. The other issue was that Symbian was owned by companies that were competitors with eachother: no one wanted to share tools for developers (e.g. no common ask ) and they didn't want a common user interface
I loved my Nokia Symbian devices, and wonder what could have been if they were able to properly respond to the iPhone paradigm change. I didn’t follow Nokia to Windows phone, I bought my first iPhone instead. Didn’t have the same cool factor as the E7, E90, E71, N8, N95 and all that. Sometimes I think if I could get a device like the E71, same size and shape with the qwerty board, but it could hold unlimited text messages and had conversations like iOS messages that I would use and love that thing. I liked how I could unlock the phone, start typing a name, press a button and immediately send them a text. I used to do it without looking. At this point I’m just rambling. Thanks for your perspective on Symbians issues.
@@boredandagitated in the Symbian ecosystem we already had touch screen phones such as Sony Ericsson P800 (mid 2002) with full screen, handwriting recognition, icons on home screen, etc. You still needed a "stylus" to interact with the screen, but we were very close. There were plans for a phone that you could operate with your fingertips like iPhone. However the phone manufacturers didn't want to go full smartphones, believing that the "phone" part was more important than the "smart" part. :( Nokia was adamant that touch screen was a gimmick, and that people wanted the S60... And naturally no sharing of sdk and compatibility. 5 years later Apple arrived and proved them wrong
@@Wolffjord Before android and iphone came out there were many win6.5 phones (i.e. T-Mobile "Wing" w slide out keyboard and touchscreen/stylus) I had one with a bluetooth satellite receiver in my car running TomTom Software. barely ANYONE was doing this at the time! There were so many java applications out for windows mobile 6.5 OS (or whatever it was) right before iphone came along and killed it all. This was same time when BlackBerrys had long rocketed to success
@@_Agent_86 repressed traumatic memories :) the prefix E was for all variable that were Enumerators. Symbian OS had a very strcit syntax that was aimed at disambiguating what was what. For example the most important was the suffix L for functions that could "Leave": "leave" meant that the function allocate memory and memory allocation can fail. This is a throiwback to the PDA origins and the very small amount of RAM available on typical devices. Memory management was very manual. Programming for Symbian did bring challenges similar to programming on embedded system of very old personal computers from a decade earlier. iOS and Android did bring a programming style more similar to the PC world with less contrainstraints in memory management.
You need to discuss DEC. Their single user RT-11 'OS' was where CPM and MS-DOS came from. They also had RSTS-E, timesharing OS for high end PDP-11s, and actually written in BASIC! and RSX-11 for 'real time' applications like controller systems for nuclear power stations. VAX/VMS was one of the most mature OS designs ever.
I gotta hand it to you, you've been on a roll recently. Every topic you've chosen for the past 2-3 months has intrigued me enough to click, despite knowing I don't necessarily love your content (respectfully) While seemingly a backhanded compliment - at its core it's a testament to the superb quality of topics you've selected.
You can say that we are still using a form of timesharing, in the way the system has to share resources and manage memory efficiently. The CPU rapidly switches between tasks, each one receiving a slice of time, creating tge illusion of simultaneous execution of multiple programs for users.
Same goes with hypervisors, switching between all those OS instances, which even after so many years, still blows me away. I still recall when a customer showed me their ‘new’ VMWare install, and transitioned a running OS from one server to another - I just stared thinking of the ramifications of what I had just watched. Geez, this was 15-20 years ago.
I once got to deliver a printout to Thompson at Bell Labs, I’m pretty sure it was something grep related. Yea, pathetic for a claim to fame, I know, but it was pretty damn cool.
The first computer I ever maintained as a Bell Canada field engineer was an Interdata Model 70 (which cloned the IBM360 instruction set). It directly ran application software (written by Northern Electric) to collect long-distance billing information from a number 4 toll tandem switch (electromechanical crossbar) then write it to a pair of 9-track HP-7970 tape decks. This system did not employ an OS which meant that the application software needed to directly support everything including all CPU interrupts and all I/O. Think about that for a moment: no device drivers; you need to implement all that stuff yourself. Anyone who ever worked on this system needed to memorize the 50-sequence which was "a bootstrap" routine required to load the application program from a cassette tape. The application programmer was a brilliant Egyptian woman with a degree in mathematics. (I do not think that "computer programmer" was a term being used at that time)
From 1982 to 1984 my employer had me run an EPA simulation model for chemical partitioning in defined environments. We contracted with a company that provided timeshare on a mainframe and I learned how to use job control language. Because I could set up multiple jobs and they would run overnight, I wouldn’t know the output until the next day. It was a lot of trial and error. And because billing was once a month, it turned out, I had blown the budget. I went back to gradual school, “where you gradually learn you don’t want to go to school anymore” (John Irving). There the computer facilities included a UNIX main frame, UNIX workstations, and some Apple IIs. What a change from JCL, punch cards, and tape. Needless to say, I have a long history using computers and operating systems. Many have gone by the wayside, though one is arguably my favorite as I still use a Palm today.
Why isn't there a discussion of DEC's PDP-6 and PDP-10 / DECSystem10 and the original 'Monitor', which was the first timesharing system? Later it was called TOPS-10. Very significant development just overlooked!
What a great channel this is. Such interesting work, buzzing in the background, as the world tuned into circuses. There are some very, very, very smart people out there. I am not one of them, but I was in IT for 30 years and as I did my work just observed it all grow. The mental grunt involved in all this stuff is quite astonishing.
I’ve been part of several of these stories. I started work in timesharing support on Honeywell GCOS, worked on Apple IIs, coded in PL/1 on Multics, wrote C code for MSDOs, then C++ for Windows, was present at the Symbian launch and wrote the text layout and font systems for Symbian, then later on wrote parts of the Blackberry OS. In my current project I code for Windows, Linux, Android and IOS. So I’ve had an interesting career so far.
I was always very into computers. From my earliest memories they fascinated me. One of the coolest things my dad would regale to me was his struggles of having to book time with a supercomputer at the lab to work on his PhD or how he spent thousands just to get a PC with kilobytes of RAM and it didn't even come up with a hard drive. It made me really appreciate the wild west of early computing and how lucky we are today.
When I began to study computers, in 1995, I wish ~ I SO wish, I had this video. They started to teach us about operating systems, but it was SO damn confusing. And in the middle of my course, the SCO-Linux legal debacle was playing out, being expensive, carrying the strong possibility that one party may control the rights to every working operating system, or at least, to everything that had UNIX in its parentage. Like Linux for example. Like (less directly, less obviously) the Apple desktop OS. Trying to get your head around the big picture, understand how all the parts of it fit together, and the fact they were all moving, like the logs prior to the log-jam, they're all moving downstream, bobbing around independently, bumping into each other, but could jam up at any moment ... I know I was told that DOS was basically a device driver for a floppy disk and a hard drive, and that everything else it did was just tacked on as an afterthought. When Win-95 came along, it added quite a respectable user interface, but it was still slapped over the top of DOS, which wasn't an operating system's arm-pit. I quite liked Win2k. I had been using NT4 as my daily, so ... I liked XP. In '97 or so, I discovered I could download a shareware version of VMware, and do guest operating systems. Hello RedHat. That was LONG before they floated as a company .... Linux disros became like one of those desktop toys for me, with the swinging balls. Something you poke & prod and play with. You could pass networking through, you could (the default) have the whole network stack inside the VM talk to the ISP and the Internet as an independent client ... there was a lot to play with and figure out. And IP6 is coming, which, means this IP4 and address translation and DHCP and all that shit ~ that complexity is going away. Right? That was bloody nearly 30 years ago! Did I mention log jams? Today? Linux Mint + Mate ~ very happy with my choice. If you listen to the Artificial Intelligence crowd (they’re hard to get away from) then the next development of everything, from the screensaver to the whole internet and computational landscape, is about to change. I’m pretty sure I don’t WANT a computer that has AI as any part of its operating system ~ let along the whole damn thing. I don’t know that I need a Trusted Computing Module (I don’t trust it) and I don’t know that I need a neural processing unit. That seems a bit like building a new church, by starting with a big hole, where you assemble and then cover a black mass altar, inverted cross, and then roof it over and build the nice koom-bar-ya church on top of it.
Laf, I can relate to a lot of this. One of my goals for going into electronics tech was to take the PC quarter so that I could understand wtf the ads were saying when describing the specs for ‘new’ pcs back in the mid 90s.
As an avid keyboard specialist and a "fairly" new fan of microcomputer technology, I enjoyed your in-depth computer history that very few people could ever teach.
I've been watching this, and other videos, for hours. And I am reconsidering my ad strategy on my channel. It is apparent that ads on TH-cam are normal for every three minutes of video. And that makes it clear that my channel's limited video interruption policy is archaic. I should give in to the trend and just go for every ad I can place.
2:04 ish where you file is really depends on your memory manager and your storage device(s). If your system is old enough, the bookcase analogy is pretty good (or maybe a pez dispenser for sequential storage, like tape drives). Of course this isn't the Usagi Electric channel so I guess most users will be using modern computers where everything is random access. So yeah, it's all over the place.
Fortunately most filesystems, like ext4 and xfs, don't store files as fragmented as ntfs. Of course it still can end up somewhat fragmented if you don't have sufficiently large enough blocks of free space on your sotrage-device and you store large files.
Publishing the source code of the BIOS was the best thing they did. It meant you could learn everything about the hardware by reading one book (and maybe a few chip datasheets)
This was more about the business case of the operating system rather than about the evolution of OS itself. You could have dived into the micro kernel concepts. You could have pointed out that macOS and iOS are basically Unix under the covers. Linux is basically Unix rewritten. You could have dived into Plan 9 and many other concepts and ventures of where the OS has gone and explored.
I suppose the future of the OS will be to have a personalized OS for every human that interfaces with the singularity through their brain implants, allowing the OS to use the individuals personality and understand and even anticipate what the user wants. That way, it will know I want to watch Asianometry between the hours of noon and 5pm daily, and when a new video drops, to cancel less important tasks(like working on my imaginary Phd) and streaming it to my eager brain stem, releasing serotonin in waves of euphoric joy. Or something like that, lol.
named & unnamed pipes are a type of IPC, along w/ msg queues, shared memory, synchronization and sockets. you could make a whole series on each of these types alone, let alone zooming out to explain everything else lol
Time sharing is also the basis of every visual os like windows and Mac OS. The processor is dividing it attention between the task of rendering the screen and running any program that is in the background.
A prompt grants you access to a peripheral sensor (the keyboard, usually) and gives you the opportunity to enter a command, hence the term "command prompt."
The CP/M story is out there, something about Kildall being up flying his plane when IBM visited and his wife not wanting to sign a confidentiality statement before knowing the reason for the visit.
Love these deep dives into the 1980s Halt and Catch Fire era that I remember as a child. Wondering if youve read the books Chip War and Route 128 that chronicle that era and if you plan on making more videos about this pivotal and poorly understood time in the history of computing? Thanks again for these delightful videos
I'd define OS as a foundational system that serves as a platform for (multiple) software applications. The concept of resource management in itself doesn't seem to contain the important function of an abstraction layer / common compatibility layer for developers. For example even things like Docker arguably does hardware resource management (but it wouldn't work without OSs). Hypervisors also do hardware resource management, but from what I understand we still require an OS to run applications on. Though, it sounds like early "operating systems" were more akin to virtualization layers.
Great video, great topic! Man I miss Windows 7 so much!! I love all your stuff but is there anyway you could please increase the volume of your voiceover just a little? Please? I have to turn your videos up so loud and when the TH-cam ads cut in they are blaringly loud! Please and thank you?!? Keep up the great work, you have the best topics presented in the best videos, thank you for all this information!
I grew up in Monterey and Cupertino during the PC boom (all of my family worked for Silicom Valley) and seeing all of these people and places brings me right back. I actually own a place near the birthplace of the RAMAC hard drive, IBMs San Jose lab, there’s a plaque in the parking lot. I’ve even ran into Steve Jobs and Wozniak just out and about on multiple occasions. It’s so wild to me to be watching this video and typing this out on my tiny iPad Mini which is so much more versatile and powerful than the computers from those times.
A different “John McCarthy” was my mentor while I was learning C. “If you program crashes, it’s because you’re scribbling on memory”. . . and “I LOVE pointers” Thank you, John
You took on a daunting task… Explaining the history of operating systems in 30 minutes or less… You skimmed over all of the macOS Apple, which were critical to the development of Microsoft windows… And of course, the pantheon of PC operating systems, including AmigaOS or Apple and even basic as an operating system… and a ton of players in the mini computer market like digital equipment, corporations vax VMS Still very nicely done!
Digital Equipment Corporation's Virtual Memory System (VMS) was and still is the finest operating system ever created and is still used in mission critical environments.
This vid brings back memories of the start of my tech career back in the late 90s. Having practically no real security made tech support much easier back then, lol. And other things as well.
To add a bit more detail to the early OSs, originally computers would run a single program out of cards. You would book say, an hour and show up with your cards. Say your program ran flawlessly in 25 minutes, you pack your things and leave. The computer time is wasted sitting idle for the the next 35 minutes. Or, if your program crashed, you would try to fix it and hope to run it in whatever time you had left. To address this, a batch system was implemented. You'd leave your cards with the technician and he would run load and run your program in turn along with all the other ones and print your results for you to collect. Then somebody figured they could load all the jobs at once with control cards between them to describe each one. Now there's a program reading these control cards and launching the other programs. Presto, the first OS.
Great content, great insight, great video!! Had *NO* idea Bill and MS bought DOS and it's developer...I thought they made it in house... *plus* the IBM-Gates connection being his MOTHER?! Man everything makes soooo000ooo much more sense now 🤦♂️🤣
very interesting video 10/10, although I already knew most of these things. kinda weird considering im from 05. btw did you know that before Windows Mobile 5, the os stored programs and other data in ram? flash was too expensive which makes it a pain when the battery runs out because all your data is gone then unless you have a backup to restore from lol
Seems there was another, the ATEX operating system with video cards in a PDP 11, it had memory mapped video and shared files across systems and had a email sharing between them, Used in newspapers and magazine publishers, As a system engineer I could look at someones crt across hundreds of miles, wire services flowed in and sent to who required them, also software to layout the format of that publiher. And alot more.
I also reflected upon Karpathys idea of LLM:s as operating systems in his video - it really stands out - but I can’t really see the direct parallel. What I can see is the part of systems that communicate between man and machine, the interface. That is perfectly a fit for LLM:s. Not the part that functions as glue between the machine and peripherals, not for printing stuff, or even low level tasks being performed in parallel e.g. So it is more of a potential GUI replacement or, maybe, keyboard and mouse in a current OS than a substitute. It maybe will as it have had, a close relationship with GPU. Or probably as it will be in the future: an addition to the user interface?
LLMs are great translation machines. They can be a layer above an interpreter like python, or they could be used to translate natural language into shell commands. Of course, paired with speech-to-text, they allow for a more seamless interaction between man and machine. In the near term, I don't expect them to replace GUIs but rather, they will supplement them with a LUI (Language User Interface) or AUI (Auditory User Interface).
I remember having a Symbian Nokia phone. For it's time, just as Android and iPhone were taking off, it was pretty advanced. It had a non capacative touch screen, bluetooth, wifi, and had some advanced apps like audiobook readers.
You are asking the wrong question in reference to will an LLM be an OS, it will never be a OS. Although, what it can be is an interface replaced for the CLI, or the GUI.
You skipped right over the entire minicomputer era and the multi-tasking operating systems developed during that era. IBM's MPX on the 1800 real time process control computer was one of the first commercial (non-military, or space) inspirations for these operating systems. Digital Equipment Corp was second only to IBM in size, and had several operating systems including RT-11, RSX-11D, RSX-11M, RSTS, Vax/VMS, etc. Data General, Prime, General Automation, and several other minicomputer companies of that era created their own operating systems. Far more systems were installed with these operating systems than Unix until the era of Sun Microsystems. So you skipped over two decades of important development in the OS history.
The Palm Pilot was a genius device - no handrwitten recognition and all the rest. I miss an app for smartphones that do all those things the Newton promissed. Note: The Apple IIgs ran the first full color windows GUI and Gary Kildall's GEM was an excellent GUI for PC and other machines. I have GEM installed on Virtualbox to play around with it.
This is every engineer's bedtime story
I'm basically using it as one right now...
I had to stop listening to Asianometry while working for this precise reason, it’s an absolute knock out for sleep deprived brains…
Or aspiring engineers.
I fell asleep to it last night lmao
Yup, every engineer who steal other's idea.
In September 1968 I was sat at a Teletype terminal creating print and 8-hole tape copies of an Algol program for an Eliot something computer in the room next door. A year later I was using a 80 column card device to produce Fortran code to run on a Univac batch processing machine that was a 45 minute train ride away from the office. No mention of operating systems yet.
My first OS encounter was on a Ferranti computer controlling a nuclear power station - an interrupt driven system using a physically huge (wardrobe size cabinet) drum for secondary memory.
That was the foundation for working with desktop Personal Computers from 1977 and following all these names like CP/M, UCSD p-system, PCOS, MS-DOS, Windows, and UNIX (several flavours).
This video seemed to map my leaning curve over almost 60 years in computing - as @KangJangkrik wrote 3 hours ago an engineer's bedtime story.
Thank you for this rewind.
I can relate. I started programming as a high-school sophomore in 1976. After passing the first course writing programs in BASIC on a Model 33 Teletype with paper tape storage I was the only student that year to learn FORTRAN. That involved going to the school district administrative building to use their card punch, then taking my card deck to the data center, and finally picking up the results the next day.
@@KurtisRader a co-ed of mine shuffled my card deck unnoticed ...
@@kyriosity-at-github
Re 'shuffling'...
...and that is why many languages and their compilers in the punched-card era had optional line numbers to allow manual sorting.
That said, having access to either a card-punch like the glorious IBM 029 which could print the contents on the top edge, or a dedicated 'interpreter' machine was a luxury to us progammers who usually "made do" by drawing a diagonal line across the top of the card deck with a marker pen.
No - we didn't use sealing wax.
@@gdm2417 yeaps, i could restore the order, but it took at least one day to receive the compiler error
Just think about this for a moment... all this has happened in less than a hundred years, from the first telegraph to having pocket computers and phones in the palm of our hand, which interface with AI/LLM's and who knows what more to come.
Never had I thought I'd hear about the human centipede on this channel, let alone using that as an analogy to Unix pipes. You've outdone yourself there, Jon.
I suspect that analogy has some staying power, since it does render the idea both rather accurately as well as in a funny, easy understand and visualize way.
@@noth606 staying power as in the mouthful of lunch which burst back out at that instant... and shall be stuck on the wall, drying up....
Looooolllllll
10 years ago, I would have been angry about this infohazard.
These days, I just chuckle.
All hail the antimemetics division!
😅
Yes!
Definition of Operating System:
"Abstracting away the horrors of hardware"
Replacing with the horrors of software.
The story of the use of computers in general is layers of abstraction. Abstraction on top of abstraction on top of abstraction, each layer allowing faster development, though also adding a layer of overhead.
There's a *lot* of overhead. Finding ways to reduce that overhead will help mitigate the end of Moore's law.
@@brodriguez11000 We have lived to see man made horrors beyond our comprehension
@@adissentingopinion848 now now enough about windows.
@@brodriguez11000 I have a few minor contributions in the Linux kernel. Hardware is outright hostile, and debugging is frustration as an olympic disciplin. Software, while as error prone as math, has incredibly powerful development and debugging tools. The problem just is that if something is comparatively easy, humans push the envelope until it becomes hard.
That grass, those rolling hills, those clouds. A little bit of me was home and back in a more innocent time looking at that.
@@merlinemeresk412 no doubt but to me it was the comfy home screen of many many happy hours learning and playing.
windows XP, what a mood.
I was just reading, that was and still is the most viewed picture in the world. Pretty humbling.
Imagine few hundred years later, a museum will play these videos to tell the early days of computer history .
More likely it’ll be, “all we know is they went digital. Unfortunately nothing survived”
In about 125 years, it'll be "You mean to tell us we're still using UNIX, a 175 year old OS?"
@@bobweiram6321
I like your version of the future better than @_Agent_86 's. 🙂
Correct.
@@bobweiram6321that’s the truth 😂
10 print "obscene word"
20 goto 10
run
My first program for BASIC when I was 7. I was very impressed the first time I saw windows and a mouse as a kid. I tried to explain it to dad at supper; he didn't pay any attention.
P*NIS
P*NIS
P*NIS
P*NIS
P*NIS
P*NIS
P*NIS
BREAK IN 10
READY.
So how did it end? Are you currently in IT?
Did that mean you created an infinite loop ?
Darned BASIC "obscene word" viruses! ;^[}
@gafakyusef6201 No, Industrial Electrician, Automation, Refrigeration, and HVAC. Programming a PLC for automation is similar to basic; especially if it is a Beckhoff twin cat and written in script.
I prefer ladder logic and Rockwell Automation with Allen-Bradley plc; ladder reminds me of basic.
I wrote many HMI programs to tie to my PLC programs; I think learning basic as a kid made it easier for me than for the older generation, when automation became more common.
"Its like the human centipede of computer processes"
"... written in the high-level C language"
"So there was Kildal, in his room, with just a naked floppy drive"
Goddamn our boy comin in HOT
Not sure what your issue is with "... written in the high-level C language", given that at the time most of the stuff under the hood was written in assembly. While "C" is considered pretty low level now, it was not at the time.
@@thekinginyellow1744 As you say, it's probably the historical irony of how abstract and high-level C was viewed at the time, compared to its current view as being merely a step above Assembler.
Ultimate Garbage In -> Garbage Out....
The "in his room with a naked floppy drive" had me spitting out my coffee
@@watchm4ker I wouldn't be surprised if all of today's high level language are gonna be seen as low level in a few years, replaced by programming through AI
In an alternate universe, the asianometry dude is the Matrix's Architect.
"Alternate" 😉
And all is well there
Yep
hahahaha
"I am the Architect. But please, call me Larry."
great profile pic, Wilco!
Retired electronics eng here. There was a story that at one point when MS was trying to kill Lotus 1-2-3 that the OS team had a saying for the next version of DOS - "DOS aint done til Lotus wont run". But that was by far not the worst thing that MS did. Often they would come to companies that had software like a TCPIP stack (I worked for a company with the first commercial stack for the PC) and say we want to license it for 20 cents a copy (we got $200). They would say either accept it or we develop it and include it in the OS. This was how much of windows utilities originated - disk utilities, FAX software, TCPIP stack etc. One by one they got put under by the juggernaut of MS. It meant more for less for consumers but often it also meant an inferior product. I can remember a large airline begging with us to not stop supporting our windows TCPIP stack because the MS stack just didnt work in their environment. BTW we also had the first browser for a PC called Emissary and its logo was a blue e - sound familiar?
You should consider writing a book. There are so many out there about the broad strokes of the PC history, I’d think there’d be a enthusiastic interest in one ff stories like yours, told from the perspective of someone who lived it.
You have taught me more IT stuff / history than I learned in my first year of Electrical Engineering. Thank you.
I wish my Samsung phone had my old Palm's Graffiti.
I am waiting on the Plan 9 from Bell Labs story.
Yesss!
They told me it exists but all I saw was ideas, and not very smart ones.
I think Raymond said it best. It was trying to be the perfect solution to a problem no one had. Unix already existed, and had moved well past those ideas.
@@YanestraAgainit exists and you can download it and install it.
@@YanestraAgainbelieve it or not plan 9 is used for wsl for its protocols involving invoking Linux and windows integrations with the file system and io
I was Superman when writing a big, functional MS-DOS batch file in early 90's. I know this may be silly, but to me personally, that was the joy of computing. Thanks for this amazing video, mister.
Love your channel. When YT feels like it's getting dumber, I'm happy to find your insightful videos.
The amount of knowledge this channel provides for free is insane. I'm definitely subscribing next month for patreon. Thank you for hard amount of research and effort into these videos
Well done John in making the distinction between operating systems and IPLs, kernel-mode programs, and HALs. Many operating systems, antique and modern, don't fit directly into any of those boxes.
Well done also for picking points that most people will connect with while still keeping it to 30 minutes, too. This could easily be a 20 part series if you wanted it to.
I particularly like that you've addressed the economic impact of the "IBM PC" marketing. The rise of the PC in the face of cheaper and more powerful options has always puzzled me.
I think Tanenbaums definition is the best (he probably didn't come up with it, but it's what he states in one of his books.) He says that an operating system perfoms two functions. One is resource management and the other is to provide an abstraction layer. A sort of extended machine. If you use the definition most people use (erroneously in my opinion), then you'll end up having to argue that Edge is part of the Windows OS. Which is pretty ridiculous. It's just a program shipped with the OS.
It reminds me of an exercise problem in the Silberschatz / Operating System Concept book. (Paraphrasing more or less) Who decides which is and is not part of an operating system: the user, the experts, or the court system?
I argue that the browser has become its own OS, as are cloud-provider-level microservice aggregates. The only limit to the OS turtle-stacking is theoretical.
@@JohnnieWalkerGreenI think the experts. Most users are quite uninformed and probably most judges too. The idea that Linux or Windows is just everything that comes with an install (in terms of calling them OSs) is very nebulous and reductive. I think most people use such a definition because they don't know any better and then if they ever have it pointed out to them that there is a stricter definition (I'd also say arguably more correct and useful) they don't want to accept or consider it because they didn't come across the idea on their own. Obviously that's just my opinion, but I do think it's not a very good definition. I mean if they remove Word pad from Windows, but keep everything else the same is it then a whole different version of the operating system? I wouldn't say so at all. Word Pad is just a user space program that shipps (or shipped) with Windows. Also the definition I gave is an actual definition. What would you say is the definition of an OS that most people use. Yes you know it when you see it (sort of), but can you really describe it succinctly and clearly. Where as if it's just some low level software that provides two distinct but useful functions that basically any modern system that people would refer to as an OS provides at it's core, then it's relatively easy to define. Anyway I know that was a bit of a rant 😅.
@@poofygoofI don't think so. It's similar to when people say Emacs is like an OS. It's just an interpreter. Browser are similar to OSs in some ways, but I wouldn't go as far as to say that they ARE OSs.
@@hamesparde9888 the distinction is arbitrary -- what makes a LISP machine from the 80s an OS but EMACS LISP not? DOS and CP/M didn't have much in the way of resource management, but don't they count as OSes?
I'm sorry that the Commodore AmigaOS gets left out of these conversations. It's preemptive multitasking would have fit in to this video well. Excellent video, as usual!
Except that it was attempting to copy Unix.... So, it's covered.
There's nothing interesting or unique about AmigaOS. It was just a barebones OS with mediocre multitasking.
@@Longlius sacrilege! I did more in 8mb than any mac or pc of the same era... add arexx for interoperable software hooks, and it was like having a superpower across my little renderfarm. I even ran photoshop in emulation faster and more efficiently than the expensive mac quadra with insane ram. For a time, it was the absolute best for multiple pro use cases.
@@briancase6180 That's valid, it was based on unix. And the Amiga's that shipped with unix was using a port of AT&T system V. Sun microsystems and Unix international used to show on Amiga's at trade shows.
@@alpaykasal2902 I think @Longlius must have been an Atari ST user. I don't think he will ever get over how superior the Amiga was at the time.
Symbian OS was born from the PDA world, focused on optimising limited hardware resources.
Programming on it had a very steep learning curve, due to this optimization and the absolute difference from typical PC programming.
It was very hard for any programmer not experienced with Symbian to move to it and port any of the existing software.
No matter what we did to improve the tools, it was hard to program.
The other issue was that Symbian was owned by companies that were competitors with eachother: no one wanted to share tools for developers (e.g. no common ask
) and they didn't want a common user interface
I loved my Nokia Symbian devices, and wonder what could have been if they were able to properly respond to the iPhone paradigm change. I didn’t follow Nokia to Windows phone, I bought my first iPhone instead. Didn’t have the same cool factor as the E7, E90, E71, N8, N95 and all that.
Sometimes I think if I could get a device like the E71, same size and shape with the qwerty board, but it could hold unlimited text messages and had conversations like iOS messages that I would use and love that thing. I liked how I could unlock the phone, start typing a name, press a button and immediately send them a text. I used to do it without looking.
At this point I’m just rambling. Thanks for your perspective on Symbians issues.
@@boredandagitated in the Symbian ecosystem we already had touch screen phones such as Sony Ericsson P800 (mid 2002) with full screen, handwriting recognition, icons on home screen, etc.
You still needed a "stylus" to interact with the screen, but we were very close.
There were plans for a phone that you could operate with your fingertips like iPhone. However the phone manufacturers didn't want to go full smartphones, believing that the "phone" part was more important than the "smart" part. :(
Nokia was adamant that touch screen was a gimmick, and that people wanted the S60... And naturally no sharing of sdk and compatibility.
5 years later Apple arrived and proved them wrong
@@Wolffjord Before android and iphone came out there were many win6.5 phones (i.e. T-Mobile "Wing" w slide out keyboard and touchscreen/stylus) I had one with a bluetooth satellite receiver in my car running TomTom Software. barely ANYONE was doing this at the time! There were so many java applications out for windows mobile 6.5 OS (or whatever it was) right before iphone came along and killed it all. This was same time when BlackBerrys had long rocketed to success
It sure was. Iirc you could use QT for the UI, but every OS interaction was weird and was prefixed E_ I’ve blocked the experience I think!
@@_Agent_86 repressed traumatic memories :) the prefix E was for all variable that were Enumerators. Symbian OS had a very strcit syntax that was aimed at disambiguating what was what. For example the most important was the suffix L for functions that could "Leave": "leave" meant that the function allocate memory and memory allocation can fail.
This is a throiwback to the PDA origins and the very small amount of RAM available on typical devices. Memory management was very manual.
Programming for Symbian did bring challenges similar to programming on embedded system of very old personal computers from a decade earlier.
iOS and Android did bring a programming style more similar to the PC world with less contrainstraints in memory management.
You need to discuss DEC. Their single user RT-11 'OS' was where CPM and MS-DOS came from. They also had RSTS-E, timesharing OS for high end PDP-11s, and actually written in BASIC! and RSX-11 for 'real time' applications like controller systems for nuclear power stations. VAX/VMS was one of the most mature OS designs ever.
And you can get VMS still today on x86 systems. Legacy systems don't die easily.
dec ftw
ima use "the sun doesn't shine on the same dog's butt everyday" phrase from now on.
my favorite video yet! keep up the good work, yo!
I gotta hand it to you, you've been on a roll recently.
Every topic you've chosen for the past 2-3 months has intrigued me enough to click, despite knowing I don't necessarily love your content (respectfully)
While seemingly a backhanded compliment - at its core it's a testament to the superb quality of topics you've selected.
You can say that we are still using a form of timesharing, in the way the system has to share resources and manage memory efficiently. The CPU rapidly switches between tasks, each one receiving a slice of time, creating tge illusion of simultaneous execution of multiple programs for users.
Same goes with hypervisors, switching between all those OS instances, which even after so many years, still blows me away. I still recall when a customer showed me their ‘new’ VMWare install, and transitioned a running OS from one server to another - I just stared thinking of the ramifications of what I had just watched. Geez, this was 15-20 years ago.
Then the Gods (Dennis Ritchie and Ken Thompson) gave us Unix and C, and it was good.
Nothing better has ever been made
I once got to deliver a printout to Thompson at Bell Labs, I’m pretty sure it was something grep related.
Yea, pathetic for a claim to fame, I know, but it was pretty damn cool.
The first computer I ever maintained as a Bell Canada field engineer was an Interdata Model 70 (which cloned the IBM360 instruction set). It directly ran application software (written by Northern Electric) to collect long-distance billing information from a number 4 toll tandem switch (electromechanical crossbar) then write it to a pair of 9-track HP-7970 tape decks. This system did not employ an OS which meant that the application software needed to directly support everything including all CPU interrupts and all I/O. Think about that for a moment: no device drivers; you need to implement all that stuff yourself. Anyone who ever worked on this system needed to memorize the 50-sequence which was "a bootstrap" routine required to load the application program from a cassette tape. The application programmer was a brilliant Egyptian woman with a degree in mathematics. (I do not think that "computer programmer" was a term being used at that time)
Whenever I see the mighty VAX mentioned in a documentary I always smile.
Man….where in the world did that sunshine on a dog’s butt saying come from?
It's been around, yet my boss's boss also had a handy one: something was "as plain as the ass on a goat." An Oklahoma guy! 🤠✌️
From 1982 to 1984 my employer had me run an EPA simulation model for chemical partitioning in defined environments. We contracted with a company that provided timeshare on a mainframe and I learned how to use job control language. Because I could set up multiple jobs and they would run overnight, I wouldn’t know the output until the next day. It was a lot of trial and error. And because billing was once a month, it turned out, I had blown the budget.
I went back to gradual school, “where you gradually learn you don’t want to go to school anymore” (John Irving). There the computer facilities included a UNIX main frame, UNIX workstations, and some Apple IIs. What a change from JCL, punch cards, and tape.
Needless to say, I have a long history using computers and operating systems. Many have gone by the wayside, though one is arguably my favorite as I still use a Palm today.
Why isn't there a discussion of DEC's PDP-6 and PDP-10 / DECSystem10 and the original 'Monitor', which was the first timesharing system? Later it was called TOPS-10. Very significant development just overlooked!
Along with TENEX and TOPS-20
Just wanna say your content is amazing. The topics and the execution are top class. Appreciate the work you do.
What a great channel this is. Such interesting work, buzzing in the background, as the world tuned into circuses. There are some very, very, very smart people out there. I am not one of them, but I was in IT for 30 years and as I did my work just observed it all grow. The mental grunt involved in all this stuff is quite astonishing.
The "Human Centipede" mention immediately brought "garbage in, garbage out" to mind! 😞
I’ve been part of several of these stories. I started work in timesharing support on Honeywell GCOS, worked on Apple IIs, coded in PL/1 on Multics, wrote C code for MSDOs, then C++ for Windows, was present at the Symbian launch and wrote the text layout and font systems for Symbian, then later on wrote parts of the Blackberry OS. In my current project I code for Windows, Linux, Android and IOS. So I’ve had an interesting career so far.
there are numerous youtube videos to source from such a life!
I was always very into computers. From my earliest memories they fascinated me. One of the coolest things my dad would regale to me was his struggles of having to book time with a supercomputer at the lab to work on his PhD or how he spent thousands just to get a PC with kilobytes of RAM and it didn't even come up with a hard drive. It made me really appreciate the wild west of early computing and how lucky we are today.
As an graduated electrical engineer with specialization in computers and vlsi your videos topics are so fascinating can't wait for new video releases
When I began to study computers, in 1995, I wish ~ I SO wish, I had this video. They started to teach us about operating systems, but it was SO damn confusing. And in the middle of my course, the SCO-Linux legal debacle was playing out, being expensive, carrying the strong possibility that one party may control the rights to every working operating system, or at least, to everything that had UNIX in its parentage. Like Linux for example. Like (less directly, less obviously) the Apple desktop OS.
Trying to get your head around the big picture, understand how all the parts of it fit together, and the fact they were all moving, like the logs prior to the log-jam, they're all moving downstream, bobbing around independently, bumping into each other, but could jam up at any moment ...
I know I was told that DOS was basically a device driver for a floppy disk and a hard drive, and that everything else it did was just tacked on as an afterthought.
When Win-95 came along, it added quite a respectable user interface, but it was still slapped over the top of DOS, which wasn't an operating system's arm-pit.
I quite liked Win2k. I had been using NT4 as my daily, so ... I liked XP.
In '97 or so, I discovered I could download a shareware version of VMware, and do guest operating systems. Hello RedHat. That was LONG before they floated as a company ....
Linux disros became like one of those desktop toys for me, with the swinging balls. Something you poke & prod and play with. You could pass networking through, you could (the default) have the whole network stack inside the VM talk to the ISP and the Internet as an independent client ... there was a lot to play with and figure out. And IP6 is coming, which, means this IP4 and address translation and DHCP and all that shit ~ that complexity is going away. Right?
That was bloody nearly 30 years ago!
Did I mention log jams?
Today? Linux Mint + Mate ~ very happy with my choice.
If you listen to the Artificial Intelligence crowd (they’re hard to get away from) then the next development of everything, from the screensaver to the whole internet and computational landscape, is about to change. I’m pretty sure I don’t WANT a computer that has AI as any part of its operating system ~ let along the whole damn thing.
I don’t know that I need a Trusted Computing Module (I don’t trust it) and I don’t know that I need a neural processing unit. That seems a bit like building a new church, by starting with a big hole, where you assemble and then cover a black mass altar, inverted cross, and then roof it over and build the nice koom-bar-ya church on top of it.
Laf, I can relate to a lot of this. One of my goals for going into electronics tech was to take the PC quarter so that I could understand wtf the ads were saying when describing the specs for ‘new’ pcs back in the mid 90s.
Thanks!
Thanks
As an avid keyboard specialist and a "fairly" new fan of microcomputer technology, I enjoyed your in-depth computer history that very few people could ever teach.
As a computer museum volunteer, I deeply appreciate your videos. The educational and historical value is excellent. Thanks so much!
I've been watching this, and other videos, for hours. And I am reconsidering my ad strategy on my channel. It is apparent that ads on TH-cam are normal for every three minutes of video. And that makes it clear that my channel's limited video interruption policy is archaic. I should give in to the trend and just go for every ad I can place.
The last Multics site to shut down was at a Canadian Forces base here in Halifax.
2:04 ish where you file is really depends on your memory manager and your storage device(s). If your system is old enough, the bookcase analogy is pretty good (or maybe a pez dispenser for sequential storage, like tape drives). Of course this isn't the Usagi Electric channel so I guess most users will be using modern computers where everything is random access. So yeah, it's all over the place.
Fortunately most filesystems, like ext4 and xfs, don't store files as fragmented as ntfs. Of course it still can end up somewhat fragmented if you don't have sufficiently large enough blocks of free space on your sotrage-device and you store large files.
Regretfully, only watched this video 6mos after release. One of your most elegant and insightful videos. Well done.
Publishing the source code of the BIOS was the best thing they did. It meant you could learn everything about the hardware by reading one book (and maybe a few chip datasheets)
For that time period it was being provided with the road, and one had to build their own vehicles.
This was more about the business case of the operating system rather than about the evolution of OS itself. You could have dived into the micro kernel concepts. You could have pointed out that macOS and iOS are basically Unix under the covers. Linux is basically Unix rewritten. You could have dived into Plan 9 and many other concepts and ventures of where the OS has gone and explored.
Prime Computer's Primos operating system should have been included. It was based directly on MULTICS and was much more robust than Unix.
Gassssp!
Imo the OS is a HAL and IO/resource management. Everything else are tools or UI.
I suppose the future of the OS will be to have a personalized OS for every human that interfaces with the singularity through their brain implants, allowing the OS to use the individuals personality and understand and even anticipate what the user wants. That way, it will know I want to watch Asianometry between the hours of noon and 5pm daily, and when a new video drops, to cancel less important tasks(like working on my imaginary Phd) and streaming it to my eager brain stem, releasing serotonin in waves of euphoric joy. Or something like that, lol.
Thank you for the great video. TH-cam is finally putting your video drops front and center! Keep up that amazing work and thank you again.
named & unnamed pipes are a type of IPC, along w/ msg queues, shared memory, synchronization and sockets. you could make a whole series on each of these types alone, let alone zooming out to explain everything else lol
I already want a video on the SAGE system.
excellent video.
I've always wondered what happened to Gary Kildall after missing the fabled opportunity.
10:11 what kind of movies are you watching?
Xxx
Time sharing is also the basis of every visual os like windows and Mac OS. The processor is dividing it attention between the task of rendering the screen and running any program that is in the background.
A prompt grants you access to a peripheral sensor (the keyboard, usually) and gives you the opportunity to enter a command, hence the term "command prompt."
Great video 👍. So nice to hear the history of a particular technology.
The CP/M story is out there, something about Kildall being up flying his plane when IBM visited and his wife not wanting to sign a confidentiality statement before knowing the reason for the visit.
Didn’t the IBMers leave then? And go see Gates?
Love these deep dives into the 1980s Halt and Catch Fire era that I remember as a child. Wondering if youve read the books Chip War and Route 128 that chronicle that era and if you plan on making more videos about this pivotal and poorly understood time in the history of computing? Thanks again for these delightful videos
I'd define OS as a foundational system that serves as a platform for (multiple) software applications. The concept of resource management in itself doesn't seem to contain the important function of an abstraction layer / common compatibility layer for developers. For example even things like Docker arguably does hardware resource management (but it wouldn't work without OSs). Hypervisors also do hardware resource management, but from what I understand we still require an OS to run applications on. Though, it sounds like early "operating systems" were more akin to virtualization layers.
Great video, great topic! Man I miss Windows 7 so much!!
I love all your stuff but is there anyway you could please increase the volume of your voiceover just a little? Please?
I have to turn your videos up so loud and when the TH-cam ads cut in they are blaringly loud!
Please and thank you?!? Keep up the great work, you have the best topics presented in the best videos, thank you for all this information!
I miss ms-dos 2.11 😉😁
I listen to his videos at full volume a level that I never get even when listening to music
10:10 Thanks for that imagery, I'll never think of process pipes the same way again.
Fabulous job! This video is like a journey through my entire career.
I grew up in Monterey and Cupertino during the PC boom (all of my family worked for Silicom Valley) and seeing all of these people and places brings me right back. I actually own a place near the birthplace of the RAMAC hard drive, IBMs San Jose lab, there’s a plaque in the parking lot. I’ve even ran into Steve Jobs and Wozniak just out and about on multiple occasions.
It’s so wild to me to be watching this video and typing this out on my tiny iPad Mini which is so much more versatile and powerful than the computers from those times.
Love this comment “… and it somehow works” (when referring to working with abstractions on top of more abstractions).
A different “John McCarthy” was my mentor while I was learning C.
“If you program crashes, it’s because you’re scribbling on memory”. . . and “I LOVE pointers”
Thank you, John
You took on a daunting task… Explaining the history of operating systems in 30 minutes or less… You skimmed over all of the macOS Apple, which were critical to the development of Microsoft windows… And of course, the pantheon of PC operating systems, including AmigaOS or Apple and even basic as an operating system… and a ton of players in the mini computer market like digital equipment, corporations vax VMS Still very nicely done!
Digital Equipment Corporation's Virtual Memory System (VMS) was and still is the finest operating system ever created and is still used in mission critical environments.
And you can get VMS still today on x86 systems.
10:17 That VAX 11/780 did not run a version of VMS that supported pipelines, and Ultrix installed on those machines were next to 0
This vid brings back memories of the start of my tech career back in the late 90s. Having practically no real security made tech support much easier back then, lol. And other things as well.
I've been married to a software engineer since 1970. This video taught me so much I never understood about his work. Thank you!
To add a bit more detail to the early OSs, originally computers would run a single program out of cards. You would book say, an hour and show up with your cards. Say your program ran flawlessly in 25 minutes, you pack your things and leave. The computer time is wasted sitting idle for the the next 35 minutes. Or, if your program crashed, you would try to fix it and hope to run it in whatever time you had left.
To address this, a batch system was implemented. You'd leave your cards with the technician and he would run load and run your program in turn along with all the other ones and print your results for you to collect. Then somebody figured they could load all the jobs at once with control cards between them to describe each one. Now there's a program reading these control cards and launching the other programs. Presto, the first OS.
12:06 that "Host Computer" is a SUN E6500/5500 connected to a E4500 and ran Solaris nothing to do with TSOS
cool video . lots to explore in future videos on this topic
13:48 Disk operating system, or DOS
I feel so enlightened all of a sudden
It just makes sense
Thanks for the great content as usual
Should’ve been titled “The Evolution of the Worst Operating System” since it was so heavily focused on MS.
Still have the Coherent for 286 computers manual from 1990. Couldn't afford a SCO Unix license back then.
Great content, great insight, great video!! Had *NO* idea Bill and MS bought DOS and it's developer...I thought they made it in house... *plus* the IBM-Gates connection being his MOTHER?! Man everything makes soooo000ooo much more sense now 🤦♂️🤣
Yes it does lol I knew they bought dos but I had no idea about his ibm connections
very interesting video 10/10, although I already knew most of these things. kinda weird considering im from 05. btw did you know that before Windows Mobile 5, the os stored programs and other data in ram? flash was too expensive which makes it a pain when the battery runs out because all your data is gone then unless you have a backup to restore from lol
Awesome Naval Postgraduate School (NPS) shoutout! Richard Hamming also worked there.
According to Gary Killdal, he did meet ibm, but there were some complications. There was an issue with the nda, and a modified version was signed.
I'm impressed, very concisely delivered. Bravo!
You do produce some very good and well varied content .. thanks ! :-)
Seems there was another, the ATEX operating system with video cards in a PDP 11, it had memory mapped video and shared files across systems and had a email sharing between them, Used in newspapers and magazine publishers, As a system engineer I could look at someones crt across hundreds of miles, wire services flowed in and sent to who required them, also software to layout the format of that publiher. And alot more.
A small correction. The batch commuting you discuss around 3:30 was for the IBM 704, not the 701.
I also reflected upon Karpathys idea of LLM:s as operating systems in his video - it really stands out - but I can’t really see the direct parallel. What I can see is the part of systems that communicate between man and machine, the interface. That is perfectly a fit for LLM:s. Not the part that functions as glue between the machine and peripherals, not for printing stuff, or even low level tasks being performed in parallel e.g. So it is more of a potential GUI replacement or, maybe, keyboard and mouse in a current OS than a substitute. It maybe will as it have had, a close relationship with GPU. Or probably as it will be in the future: an addition to the user interface?
The LLM may have a deeper understanding of the abstractions below it. The better to getting the most out of them.
LLMs are great translation machines. They can be a layer above an interpreter like python, or they could be used to translate natural language into shell commands. Of course, paired with speech-to-text, they allow for a more seamless interaction between man and machine. In the near term, I don't expect them to replace GUIs but rather, they will supplement them with a LUI (Language User Interface) or AUI (Auditory User Interface).
I suppose you didn't have space to mention TSS/8 and TOPS-10, both pretty elegant.
I remember having a Symbian Nokia phone. For it's time, just as Android and iPhone were taking off, it was pretty advanced. It had a non capacative touch screen, bluetooth, wifi, and had some advanced apps like audiobook readers.
I don’t quite understand using split eq on material that doesn’t have transients. Isn’t this basically multi band mid/side processing?
I'm here after one minute, and I'm still disappointed I wasn't sooner.
Thanks for the trip down Memory lane
no mention of amiga workbench or GEM ? or even amiga multitasking first in the world on the pc thanks to dedicated chips.
You are asking the wrong question in reference to will an LLM be an OS, it will never be a OS. Although, what it can be is an interface replaced for the CLI, or the GUI.
You skipped right over the entire minicomputer era and the multi-tasking operating systems developed during that era. IBM's MPX on the 1800 real time process control computer was one of the first commercial (non-military, or space) inspirations for these operating systems. Digital Equipment Corp was second only to IBM in size, and had several operating systems including RT-11, RSX-11D, RSX-11M, RSTS, Vax/VMS, etc. Data General, Prime, General Automation, and several other minicomputer companies of that era created their own operating systems. Far more systems were installed with these operating systems than Unix until the era of Sun Microsystems. So you skipped over two decades of important development in the OS history.
Ok. Well, good thing you commented then.
The Palm Pilot was a genius device - no handrwitten recognition and all the rest. I miss an app for smartphones that do all those things the Newton promissed.
Note: The Apple IIgs ran the first full color windows GUI and Gary Kildall's GEM was an excellent GUI for PC and other machines. I have GEM installed on Virtualbox to play around with it.
I once described piping to a coworker as the human centipede. The description here made me happy.
thorough as always, thank you
So much love for the tech expressed in the video ♥