My first experience with Unix was minix in 1990. Before that I was a TOS and DOS guy - always disliked DOS. I was immediate sold to the basic Unix philosophy of “everything is a file”. I find that Linux unfortunately is slowly moving away from that paradigm. I started Linux in 1995, at the same time I had an Indy with Irix. Irix is a nice clean SystemV Unix. Then in 1997 I got a written off, microPDP11 from my work. I put SystemIII on it. Was quiet an ordeal so early in the time of the internet (not a lot of downloads available). Then I did Tru64/TruCluster (the best ever Unix!), AIX, HP-UX and SunOS and Solaris. Those days of the mid 90s to the early 00s were magical for Unix. Because the internet started to happen and that was what Unix excelled on. Great standard development tools, good networking stack, lightweight. Man… miss these days.

@@CallousCoder The only proper UNIX I've ever used was Sun Microsystems' Solaris. I think Sun made some amazing products. I have a little bit of Linux experience, but I prefer FreeBSD to Linux in most cases. I agree the late 90s to early 00s was peak for Linux (as well as UNIX) and whatever Linux is these days pales in comparison to that era. I miss those days too man...what an exciting time it was!

@@toonz9971 then you haven’t used Irix or Tru64. Full SystemV systems and DEC Tru64 with its Alpha blew everything out of the water with its speed. And then adding TruCluster to the mix made a 24x7 Unix you could migrate a running process to another node, not even the socket was disconnected just a little “hiccup as it moved across the memory channel”. And you could just patch the code or the OS. It was marvelous! And their SpecFp was a long-standing record. When Sun and SGI were doing 200MHz on 32 bits we were pushing 700MHz with full 64 bit. My jaw dropped back in 1995 when we gotten two GS140s with TruCluster for our 24x7 available database and fluid sim calculation. Me colleague who made the fluid sim code was like… how many CPUs 14? Oh… I think I will do my first ever parallel programming 😀And me well technically 28 but then you may want to fork a second instance. I did btw love Solaris 10 on Intel! I used the shit out of containers and zfs! I immediate realized containers solved our development system problems. As well zfs, snapshot the database run your integration tests rollback snapshot. New container for your feature, there you go! Have fun! It was amazing! We had lpar on AIX but this was better. Now Docker took its place.

My first Unix was SunOS 4.1.x in our campus computer lab in the early 1990s. SunOS was the precursor to Solaris. (Technically, SunOS 4.1.1 was also "Solaris" but it wouldn't be branded as "Solaris" until a little later than that - we ran SunOS at my first job, and Solaris at my second job.) I'll always have fond memories of running SunOS and Solaris. We had Sun 3/50 diskless workstations on campus, but I loved the SPARCstation "pizzabox" (S4? S5?) at my first job.

Glad you liked it! I always have a terminal window running on my Linux machine at home, even though I run GNOME as my desktop and run GUI apps. The terminal offers a very flexible and powerful command line.

Glad you like it! I wrote an article about the experiment of using Linux like original Unix, but "screenshots" in an article doesn't show the experience quite like a video. I don't usually use ed for big projects - but when I could only use ed for this experiment, I found I quickly adapted to using it. While working on larger editing projects (I wrote several nontrivial programs and documents that I didn't show in the video) using ed became second nature. It's just a different way of thinking when I edit. 👍

If you use sed, you're using ed commands.

Fast forward to 2023 and still no good non-interactive editor

@@OrmaajPerhaps that is because very few people would even want one.

@jnharton Non-interactive editors are useful. One of the options for the diff command (I think it's -e) makes it generate a script of ed commands to turn one file into another. This was then used to build the first version of SCCS, which later led to RCS, CVS, Git, Fossil, etc.

Pretty much, yes. Awk (1977) made it easy to write scripts to do lots of useful tasks that you'd otherwise need to write a one-off program. Sed was useful for simple mass edits, grep was excellent for one-line patterns. Awk made it easy to do things like "If *this* line in the log file says 'Error' then print the next few lines (line count might vary) until you see some end-marker." That's easy to do in awk, not possible in grep.

hi! just to confirm, u can print a predefined number of "context" lines with grep too right? the -B or -A flags. > _"[untill end-marker] That's easy to do in awk, not possible in grep."_

What a great future world you must come from! ☺ The "grep" command was added in Unix 4th Edition in November 1973 - so it wasn't available in the Unix 3rd Edition I'm showing here. And no, grep didn't support context lines in 4th Edition .. it was one line of output per match. (Basically the same as the g/re/p command in the ed editor.) I'm not sure when Unix grep supported context lines, or if that was a GNU extension. And the example I mentioned in my above comment was meant to suggest a variable number of lines .. like a multi-line error message that might end with some "end error" message on its own line (typical in some large environments I've used). That's easy to write as an awk program (if line matches, then doprint=1 .. print if doprint==1 .. if we find the "end error" message, then doprint=0).

If you can try a bit of FORTRAN, go for it. Just play with some non-trivial maths problems, and you'll be surprised at the ease it affords. I firmly believe that *all* programmers should have at least had a look at one of the early languages… it gives you an insight into how far we've come and helps you to appreciate what the modern high-level language takes care of for you.

I agree! I also started a FORTRAN 77 programming series on this channel. I've done one video so far, the next one will be this weekend.

You will use UNIX, every time you use a POSIX standard OS or a Linux distro. Many things change, but the core remains eternal

I was in the fourth grade when you were born, neat.

Modern Fortran has not been column based since the spelling stopped being in all caps (in the 80s) and it dominated HPC until just the last decade when C++ has started to overtake it. These language choices are partly due to speed/efficiency but it also has a lot to do with the availible proven debugged math libraries, no sense spending a week and $100k of electricity on a scientific calculation just to be slightly uncertain of the accuracy. The fixed-column thing was a direct artifact of punched-paper programming. Java is OK for portable widget apps that are distributed to run on many operating systems, but Java was in decline until Android came along and gave it some customers. Any language that is interpreted or JIT optimized at runtime, or even just garbage collected, is no good for applications like industrial machines, healthcare, music recording, and other real-time tasks that need deterministic behavior. (Real-time doesn't mean fast, it means the actual physical wall clock latency must be bounded with guarantees of worst case maximums and minimums.) Eg some fool at Boeing tried to use Java in a new landing gear control unit and during bench testing it would just fail to extend a few percent of the time; the reason was the JVM doing some "helpful" housekeeping thing in the background would block reception of the interupt signal. The C-17 engine contol unit on the other hand has mathmatical proofs of worst case behavior (This took several years though, it is proof of the whole hardware and software stack combined.) Simpler hardware generally helps a lot in this regard, eg an little AVR arduino can handle control of more stepper motors more accurately than an i9 desktop. The i9 has a latency variation of 150µS even after disabling hyperthreading and using a hard real-time operating system. Then you add the uncertainty of PCIe based I/O and such. This is why the 80386 remained in production through 2007 despite being of no use for new desktops after about 1993 and long after the 80486 went away. 386 was the last predictable x86 design. Newer x86 CPUs have out of order execution pipelining, branch prediction, cache effects, register renaming and so on and so forth. Even though you might know it takes 5 cycles to do a multiply on a modern CPU you can't know when it will get its timeslice or when it will flush to output.

Glad you enjoyed it! This was a fun video to do. 👍

Glad it was helpful!

And I recently interviewed Brian Kernighan about it for Technically We Write: technicallywewrite.com/2023/06/01/groff

Sounds like we are about the same age. When I was an undergrad (early 1990s) we had terminals around campus. I think they were VT52 or VT100 .. maybe both, depending on the building? I used those a lot between classes to check email or write papers. I discovered Linux🐧 in 1993, and installed it (dual boot with DOS) on my computer at home. Check out my other video about how I wrote papers using nroff - if I didn't want to reboot into DOS, I could use nroff to write my papers and print from my dot matrix printer in my dorm room, or the (free) dot matrix printer in the computer lab.

@@freedosproject I'm a bit older, if I had gone straight to college, I would have graduated in the early '80s. I had a couple friends who were into computers, including one who took home a VAX while working for the phone company. So I knew a little, but not much. Electrical engineering was my major, and I was the only one in my class who wasn't planning on doing digital and/or ICs. Back then, I thought that home computers were a novelty, but not that useful. Networking really changed that. I discovered Linux while working for another university in 1995. MS-DOS / Windows 3.11 was the standard OS, but I discovered that I could dual boot into Linux and access the files on my work computer from home. (Most network cards at the university had public IP addresses back then.) I never used nroff myself. Back then I was a big fan of plain text. I used my free class audit privelege to learn C programming, and instead of turning in my assignments on fanfold paper from the line printers, I turned in fanfold from my MX-80 hooked up to my home computer running Linux.

@@StringerNews1 Loved the Epson MX-80 dot matrix printer! We had one in the 1980s and it was a workhorse. Printed out countless code listings and class papers, and just never broke down.

@@freedosproject and when I installed OS/2 Warp on the same computer, it revealed a near-letter quality mode that was a lot more readable.

All seasoned Unix sysadmins of the latter 1980s and into the 1990s knew that the first thing you did when setting up a new box was to download and install the GNU utilities.

Glad you enjoy it!

Thanks! Glad you liked it. 😃 Most of the videos are FreeDOS or FreeDOS-related, but I like doing videos like this every once in a while.

I never used a Teletype or Telex terminal as a terminal, but the physics department in my undergrad years (I was a physics student) had a Teletype terminal set up as a printer in the mini computer "lab" (a tiny room with two VT terminals, plus a pen plotter, plus the Teletype "printer").

Glad you enjoyed it!

It has a far ranging legacy, with descendants like ex, vi, sed, grep and perl. The ex language is still there in neovim, although they removed ex mode to avoid confusion.

And you can "script" ed by doing something like this: ed

ed is pretty cool. I don't use it all the time, but I do use it occasionally when I need it. Good tool to have. And if you're like me, just to play around in sometimes.

@@freedosproject Still can, though your example is more of a sed -i job. Another relation is that diff can output ed scripts.

@@freedosproject I've just never seen it behave so well and fast, I've thought ed was a slow, grinding editor, but it's actually pretty elegant! I'll definetly try it more, I even played with it for a while after watching your video. Cheers!

Glad you like them!

😎

Indeed! In Version 6 Unix, the DESCRIPTION section of `man ed` starts with this exact sentence: “Ed is the standard text editor.”

Will do! I'll try not to do too many, since I mostly cover FreeDOS things - but I think occasional videos about other retro computing is okay too.

@@freedosproject I know that you mostly cover FreeDOS things, for that I appreciate it even more. I saw that you have a list of videos on "FreeDOS Programming" and I'm going to watch them. I really like the software development terminal session videos that explain how the development environment is built and the workflows used. Thank you very much for your content!

You're very welcome! Since doing this video, I sometimes use ed (and FreeDOS edlin) unironically to edit files. Usually if I just need to make a quick edit to a file, or if I run several commands and realize "that should be a script" .. I can do ed and edit my file while I can still see the previous commands I typed.

Glad you liked it! Using Linux at the command line like this (and using ed) is really fun, and it's an interesting experiment. When I did this experiment (to use Linux like original Unix) I lived at the terminal. I just maximized it on my system, and did everything there. [Technically I cheated .. I had my work email up on my other monitor, so I can keep an eye out for client emails.] When you write nontrivial programs using ed, your brain starts to work differently. You build a kind of mental map in your head, and you can navigate pretty easily using ed. I became quite comfortable with ed - and I sometimes still use ed if I need to write something quick .. like make a script out of hte commands I just ran.

Glad you enjoyed it! ☺

Glad you liked it! ☺

I'm glad you liked it! I think it's pretty cool too. 👍

Glad you liked it! 😎

Glad you enjoyed it!

Glad you enjoyed it!

Thank you very much!

I'm glad you liked it! This was a fun video to do. ☺

More to come! I'm doing a "learn FORTRAN77" series on the channel. I've done just one video so far, but I have the rest of the series planned out.

I will try! I'm planning more FreeDOS videos next, but I'll do another one on Unix history sometime.

@@freedosproject thank you!

You never know. I'm actually so glad that I came to the party over 30 years ago, and got to mess around with lower-level stuff. In a way, it makes you more appreciative of the tools we have now, and also know their limits. ✌

Thanks! 😀 I also recorded a followup video showing "old style" C coding. You can find that on the channel, too.

🤔I'm sure Apple II didnt have "ed". Back I. The day, I remember getting countless reminders that it was pronounced ee-dee, but DOS also had an Ed that everyone pronounced "ed", so that battle was pretty much lost.

I learned it as "ed" back in the day, but later learned it was properly pronounced "ee-dee." But old habits are had to break. 😅

Glad you enjoyed it!

Glad you enjoyed it!

We need the Czar, we need the Czar Wait, you mean a Char for the Structure? No, man, the Czar is dead, long live the Czar. (From digital archeology from a Fortran, BCPL manual, late 70s. Not from me).

Glad you liked it! I found nroff and groff was really easy to use when I first learned it as an undergrad, and I still like writing some things in groff today.

Thanks! I try to keep the channel focused on FreeDOS, but sometimes I like to show Linux/Unix history too. ☺

Glad you liked it! 🐧

Glad you enjoyed it! ☺

Thanks! I'll try to share presentations like this whenever I speak on a new topic at conferences.

Very cool! 😃 Brian has a great story about Bill Joy writing vi, in his book 'Unix: A History and a Memoir.' Short version: Bill hung around the Labs when he was a grad student and showed off this new editor he was working on. Brian told Bill that he should focus on his degree, not the editor. Brian comments in the book that maybe he shouldn't give career advice. 🤣

​@@freedosprojectexamination of 4.1c source code especially from the 4.1c.1 directory tree on the CSRG ISO shows traces of early SunOS. There was a partial implementation of a "unmount(2)" system call that was backed out. SunOS kept that system call. There is a kernel configuration file "files.sun". Guy Harris admitted on USENET in 1990 that SunOS ran on a VAX-11/750 and stated that Sun Microsystems had one of those minicomputers. I still don't know when they acquired it. There is a chance that the earliest SunOS development took place in the UCB CSRG before SMI got their own VAX-11/750.

@@freedosproject "Bill hung around the Labs" making 4.2BSD a reality while also helping himself to DoD time and materials to write his proprietary SunOS.

@@freedosproject There were holy wars fought over the best choice of editors to be used. Ed was primitive and time consuming when I first started using Unix back in 1985, then I discovered Vi, loved it. Then in 1986 GNU Emacs came along and thus began the holy wars over which was better. I tried learning Emacs and it felt too alien and not intuitive, which others strongly disagreed with. Those green screen CRTs brings back a lot of memories, first used printer terminals but they were strongly discouraged for use in programming because of the consumables cost. Then switched to VT100, and finally used Visual550s which was a larger screen and actually had a Textronics graphing mode for 2D plots! Made some amazing plots for physics assignments and could preview them on screen before sending the output to the wide carriage plotters down in computing services. A lot of memories of this era… used to program in f77 on the campus time-share Honeywell Multics mainframes, never fun when class assignments were due and typing a character sometimes took 30 seconds to come back and forget about Fortran compiling during peak times. The arrival of the first Sun-3 and 4 Unix workstations was a godsend! But only the profs with fat research grants could afford them…

Glad you like it!

I'm glad you liked it! 🤓 I think it's important for every CS student to have at least a basic understanding of how we got from "there" to "here" - that's why I did this video. You might also be interested in some of the articles in the "A Look Back" series that I write on Technically We Write: technically we write . com

vi = Visual editor, Improved. it still has an ed mode.

@@pauldwalker Technically, vi actually stands for the "visual" [mode] in Ex (not Ed). It was a shortcut for launching Ex directly into visual mode instead of line mode. Then we got "Vim" (vi improved) several years later.

@@mkd1964 i’m getting old. i could have looked it up but went with my creative interpretation instead.

@@pauldwalker You're not the only one. I thought it was feature of Ed as well, until someone recently reminded me in a different video. 🙂

Thanks 👍

Glad you liked it! 😃

No it wasn’t yet. I had to look up my first edition C book by Kernighan and it wasn’t no.

I came here to say the same thing. The first c compilers I used couldn't cope with the ANSI style function prototypes as in his example. They were a huge improvement, and came at the same as much better type awareness in the compilers. I found a couple of potential nasty bugs in my existing code when I ported it to a machine with a more modern compiler.

Oops - good catch! I'll have to do it the K&R way when I do this "for real" as a conference talk. Here's a "K&R C" version of the "66 lines" program: #include linenum(in,out) FILE* in; FILE* out; { int ltr; int line; line = 1; fputs(" 1:", out); while ( (ltr = fgetc(in)) != EOF ) { fputc(ltr, out); if (ltr == ' ') { fprintf(out, "%2d:", ++line); if (line == 66) { line = 0; } } /* if ltr */ } /* while */ } main(argc,argv) int argc; char** argv; { int i; FILE* pfile; /* usage: linenum [files..] */ for (i = 1; i < argc; i++) { pfile = fopen(argv[i], "r"); if (pfile == NULL) { fputs("cannot open file: ", stderr); fputs(argv[i], stderr); fputc(' ', stderr); } else { linenum(pfile, stdout); fclose(pfile); } } if (argc == 1) { linenum(stdin,stdout); } return 0; }

int argc is the default, so you don't have to say it. Also, argc comes first.

@@b43xoit As far as I have always seen on Unix and Windows, is that you always have to define the integer. It will not assign it implicitly.

Glad you like it! The original ed command make sense after you've used them for a bit. Funny too that you start to think a little differently, you build a sort of mental map of the file and where you are in it.

A friend of mine was born mid-March 1970, and he claims to share his b'day with Unix! 🙄 I don't think an exact date has ever been pinned down, but could be I suppose.🤷‍♂️

@@squirlmy Well, either way, I hope we can agree that Unix existed before 1993, when I first used a VT100 terminal, which would have been used early on in Unix history...

We actually connected veterinarians using dial up and a terminal to their central system. It was the most affordable way.

We had VT100's on campus - they got rid of them after I graduated. If I knew then what I know how - how "retro" cool they were - I would have found the space to keep one when the campus CS department gave them away for free.

Definitely this. I remember using top on a 9600 vt100-clone and you could clearly see the cursor flying around... I remember that more than the stereotypical "long phosphor" that things like Cool Retro Term lean heavily on. I can't figure out how you would limit the speed of a vty without an actual serial cable though. Cathode on the mac does have this option, from memory.

I still prefer CRT for certain things, like working on files. I like the throwback feel.

Thanks! I'm glad you liked it! I'm giving this as a conference talk in a few weeks. Also, I just posted a new video that's very similar to this, as a "history of tech writing" video. I know it's repeating some content from this video, but it's for a different article to celebrate National Day of Writing, so I wanted to record a shorter video just for that.

I agree! 🤩 I grew up with an Apple II (actually a clone of the II+ called the Franklin ACE 1000) back when we measured memory in a few kB. I sometimes teach a university course on the history of tech, and I try to share what that was like - and how amazing it was to experience a powerful command line operating system like Unix for the first time.

😅 You're welcome!

Slow serial link with line noise, even more fun.

Yes, UC-B is where Bill Joy wrote vi. Brian Kernighan has a note about that in his book 'Unix: A History and a Memoir.' The short version is Bill was a grad student at the time, and was working on Unix, and sometimes visited Bell Labs. During a visit, he showed off this editor he was working on .. video terminals had replaced paper terminals, and he used cursor addressing to move the cursor around. Brian said it was nice but maybe Bill should focus on his PhD. Brian says that he cites this example when students ask for career advice.

And you can bask in the warm "glow" of a simulated CRT. 😅

Glad you liked it! I really enjoyed this one. Looking forward to doing it as a live conference talk sometime soon.

Someone else pointed that out too. Good catch - I missed using K&R style in the video. I'll have to remember that when I do this as a live conference talk. ☺

Same for me, but a few years later. My first job was managing Apollo, Sun, and HP - and I added Linux. My second job had HP and AIX - plus Linux. My third job was also a mix. I didn't learn the GNU additions until much later, after I no longer was a sysadmin, no longer had to worry about "will my scripts run everywhere?"

Unix before pipes - gee, talk about some stone age computing using bear skins and stone knives!

I wrote an article once about "how to check spelling the 'old school Unix' way, but it was easier to show with pipes. Without pipes, it would have been like this: tr 'A-Z' 'a-z' < test.me > tmp1 tr -d ',.:;()?!_' < tmp1 > tmp2 tr ' ' ' ' < tmp2 > tmp3 sort < tmp3 > tmp4 uniq < tmp4 > tmp5 comm -2 -3 tmp5 words ..and then you get a list of words that are likely misspelled. (For anyone who doesn't understand the commands: convert all text from the teset.me document to lowercase, remove special characters, translate spaces to newlines, sort the output, remove duplicate lines, compare the list to show only the lines that don't appear in the 'words' dictionary.) 'sort' is from Unix 1st Ed, 'uniq' is from Unix 3rd Ed, 'comm' and 'tr' appeared in Unix 4th Ed (1973). And by 3rd Ed, you had pipes anyway. In 1st and 2nd Ed, you would have written a simple program that de-duped lines from input (basically 'uniq'), a one-off program to turn a document into lowercase words one per line, and another one-off program that compared line-by-line input to a list of words and printed the "not found" lines: ./makewords < test.me > tmp3 sort < tmp3 > tmp4 ./dedupe < tmp4 > tmp5 ./nonwords < tmp5

@@freedosproject There's a great video about UNIX from the AT&T History channel in which Brian Kernighan shows how to use pipes to do a basic spellcheck. It's so funny to watch, A) because he looks so young, and B) because he has his feet up on the desk, keyboard in lap, casually showing the viewers features of UNIX that were revolutionary. There were some other, lesser known Bell Labs employees who also did a lot for UNIX, including Lorinda Cherry, John Mashey, etc. They deserve credit too, though of course Thompson and Ritchie really did the heavy lifting for UNIX.

That's great! 😃 The first Unix system I used was on a Sun 3/50 running SunOS (version 3? or was it version 4 by then?). My first job used a mix of Unix systems, including SunOS, HP-UX, and Apollo DomainOS/AEGIS (which provided both BSD and System V) … until I replaced some of the older systems with Linux.

Yeah ahhahaahaha

I do love the command line. My Linux machine is running the GNOME desktop, but I usually have a terminal open to do command line things once in a while. I grew up with the command line, so I'm very comfortable there. 🤓

I'm using Cool Retro Term. You can find it here: github.com/Swordfish90/cool-retro-term

Oh no! 😧 I didn't realize I forgot that (27:50) Good catch!

​@@freedosproject In a way why we do free software is, because we know there are always some mistakes that slip in, but we're not alone, isn't it? ;) Thanks for all your work on FreeDOS, you guys rock!

Very cool!

Thanks Rugxulo! 👍

It really is! I use ed whenever I need to capture something quickly, especially if I need to write a quick 3-line script to process some data and I want to keep the previous output on my screen while I write the script. So easy to use ed(1) when I'm on ssh somewhere. Also, I use edlin on Linux sometimes. Gregory's FreeDOS Edlin compiles and runs fine on Linux. I should do a video about that… 😃

Now that’s an old one! I’m not even sure I know how work that right.

I first read about C and Unix around 1982-1984, just before starting in highschool. There was a public technical library at the engineering school in the town, and I read dozens of books on Unix: tutorials and introductions, programming, Douglas Comer's book on Xinu (OS design), and many other classic books. I only touched a Unix machine once in 1985, a Cromemco machine with Cromix at a computer trade fair - I dared type ls on a terminal when nobody was watching. It would not be until 1990 at uni when I finally had proper access to Unix. By then, I had memorised enough from the books to feel at home immediately, and with ed bring easier to learn "by the book", that's what I used for a long time (for big files I would edit on a Mac with BBEdit using CAP AppleShare or FTP to move files. I'd also typically login with NCSA Telnet.)

Will do! I'm back to doing more DOS videos now, but I'll do another Unix video sometime.

I use different coding environments for different things-but for most of what I do, I use Vim. I know what I'm coding and how to code it, I usually don't need the help that an IDE provides. For very simple stuff, like demo programs or just some test/scratch program to try out an idea, I'll even use Edlin on FreeDOS or Ed on Linux. ☺

Thanks! I'm doing this talk later this Fall for the VCF Online presentation series. Watch for it!

@@freedosproject I will, thanks !

Thanks - glad you liked the video! I grew up on green and amber screens, like on the Apple II.

@@freedosproject That's fantatic. I remember the Apple IIs didn't have lowercase. How far we've come! :D I honestly really like older hardware and the slow speed it ran at. It felt so cozy. You had to wait for things to happen, and had time to brew some coffee! Maybe I should run Cool Retro Term + something that limits the baud rate of the output for nostalgia. :)