Had a modem and connected to a lot of bulletin boards. I felt very much a like "WarGames" hero, calling phone numbers and all :)

I still remember the day when I used ppp to connect to the actual Internet for the first time. After a lot of trial and error, I finally pinged www.linux.org, and it resolved the hostname and returned the ping. I was silent first, and all I could say then was "whoa".

Oh fun times. :)

Back in those days multi user shell machines were quite common and some of the first ISPs offered Internet access that way.

Mine was an ISP called Internet Access Cincinnati that gave each user shell on a Sun Sparc 10 running SunOS (pre-Solaris). You also got POP e-mail and a -username home folder accessible via web.

Also getting online was one of my first motivations for using Linux, as the Trumpet winsock networking library for Windows 3.1 was terrible and crashed the entire computer all the time.

I think I heard the NetBSD toaster is out of support :(

But by the early 2000s we seemed to switch to a world where I just assumed Linux would support my new hardware, be it a used Zip-drive, a web-cam, or whatever.

These days I'm the same. I fearlessly buy random motherboards, and hardware toys, and just assume they work. So far I've been lucky, but it is a real testament to the driver-developers. The lone developer who makes one pull-request against the kernel solely to add the magic numbers to recognize yet another kind of SD-card, or similar. Really benefits us all, even if they never do another thing.

But by and large, thing did still just work.

Around that same era Windows wasn’t exactly problem free with device drivers either. A fresh install on random hardware would often require hunting down Windows drivers online. Though of course those drivers did always existed because of Windows’ market share.

Now you reminded me that I bought a wifi dongle around 2012/2013 and ended up having to return it because I couldn't get it to work. But the very next one I bought as a replacement worked first time. So perhaps I'm having rose-tinted memories!

Somehow I still refer to these as "the good old days".

Until that I'll buy AMD.

Very much so -- many drops form a torrent.

Much like the post world war civil society, economic and diplomatic development was an exception in Human history.

Assuming the reality we have will remain forever without protection or constant maintenance effort is dangerous.

It's rather the opposite, combined with moral stance of Linus and it's core committee itself that has managed to kept Linux on this track so far.

These days even GNU is a small part.

systemd, Wayland, Xorg, PulseAudio, CUPS, KDE, Gnome, etc are all maintained outside of Linus and GNUs governance.

What FOSS operating systems (and not even Linux specifically) offers is freedom to choose. But the effort of using some alternatives can be significant.

So I wouldn’t say Linux is immune to having its “well poisoned” because if Linus.

True. On the other hand, Linux (the kernel) does provide the foundation. You can replace most of components above it and you will still have hardware support for a wide range of hardware. Even if you were to develop an entirely new kernel, the Linux kernel can serve as a proxy for documentation for undocumented or poorly documented hardware.

I realize that places a disproportionate amount of emphasis on the device drivers when the kernel offers so much more and an incredible amount of effort goes into the ecosystem outside of the kernel, but the reality is that most alternative operating systems grow beyond the academic toy phase. Part of the reason is the difficulty in attracting potential users, including developers, if those users have to use said operating system in a virtual machine.

I don’t know what the driver story is like for all of the other main BSDs but DragonflyBSD is a really interesting platform with some real innovations in the kernel.

Technically XNU, the macOS kernel, is open source too. Though there doesn’t seem to be much community effort behind that.

There’s also OpenIndiana. Which itself has some really novel ideas.

Then you have smaller projects that have compatibility layers, for example Haiku can run some Linux and FreeBSD systems. ReactOS and Windows driver support. But even if you exclude them in favour of larger, more stable projects, there’s still plenty of options out there.

I do also think you’re placing far too much emphasis on the kernel here. For example some platforms don’t guarantee ABI compatibility between releases, instead requiring their standard libraries being used as the kernel entry point (eg Darwin instead of XNU, or libc on some UNIXes). This is why people often talk about GNU/Linux, Darwin or NT as more than just a kernel.

And that’s before we address the massive elephant in the room that is all the abstractions that sit on top of the userland. For example how so many applications these days are basically running inside a virtual machine called a “web browser”. These days a modern browser is almost like an OS in its own right.

Don’t get me wrong, I’m not taking anything away from Linux nor the Linux kernel. But its main strength is its momentum rather than it being unique.

For example like how OpenIndina was forked from OpenSolaris (CDDL) when Oracle dropped OpenSolaris development and support.

APSL has been approved by the FSF which other “Source Available” type licenses aren’t.

Personally though, I’d still rather people supported (for example) FreeBSD rather than Darwin/XNU.

100% agreed on BSD, though I'd opt for OpenBSD myself. Either way works!

The main issue i see is that we aren't really "cultivating" the next Linuses

(In the 1990's businesses either had to develop and maintain their own operating system, or license it from someone else. Both routes are expensive. There were also collaborative efforts for larger operating systems, none of which seemed to reach positive conclusions since the businesses who owned the rights to the operating systems were ultimately in direct competition with each other. For whatever reason, the open source collaborative model used by Linux has been successful. In all probability it is because nobody "owns" Linux in the conventional sense.)

If your fulltime job is Linux it does give you an ability to influence what happens that others who are earning their daily bread some other way lack. You can use that ability to ride over some objections.

For some people the goal of Linux is to become the world's OS. I used to think this but I realise now that if it did, then it would become static and conformist. This is because the experience of using it would have to be stable and common so that all the users who don't care about Linux per se wouldn't have to keep relearning how to use it or at least would only need to relearn 1 way. So there would have to be the world linux distro, probably run by some bullying company that would keep app developers, hardware vendors in line.

We'd be complaining about it like we might about MS or Apple. I like Linux for the wrong reason - I like it because there's so much choice. Simplicity is about eliminating choice.

This is extremely worrying. There already is enough linux monoculture in critical network infrastructure due every manufacturer except Juniper seemingly putting their penguin eggs into one basket. I can only hope there will be some common sense brewing for the future before linux experiences major failure and Pluto's Kiss starts to feel a bit too real for comfort.

We should study the social and economic dynamics and game theory of this and try to replicate it. Someone should do a whole game theory Ph.D thesis on Linux and how it manifests an un-tragedy of the commons.

Claude Shannon 2001

Edsger Dijkstra 2012

John McCarthy 2011

Niklaus Wirth 2024

Konrad Zuse 1995

Alonzo Church 1995

Jim Gray 2007/2012

Grace Hopper 1992

Even Alan Turing could still have been alive during my lifetime hadn't he been murdered.

Friedrich L. Baur, who died in 2015, still had an office at my university when I studied there and I often passed by, saw his name plate, and wondered if this was all real.

Luckily Donald Knuth is still around.

Some also believe he was poisoned.

I assume that is where the “murder” comes from.

I'm glad that's finally settled. :)

Great read!

Linux distributions like Red Hat really squashed those concerns within a short time and made things much easier on all fronts. Their documentation was also excellent.

By the late 1990s, Red Hat was as common as Solaris, AIX, and IRIX at the university, but most still considered it inferior to Solaris. And now, RHEL is the main *nix system used there (a few use Debian).

It was the XF86Config modeline HOWTO that taught me how a CRT controller works.

This sound has lived rent-free in my head since 1994:

https://en.wikipedia.org/wiki/File:Linus-linux.ogg

When I met Linus T in 1994/5 I asked why he spoke Swedish and not Finnish and he rolled his eyes. Was so sick of the question even then.

https://linuxjournal.rubdos.be/ljarchive/LJ/008/2867.html

https://www.amazon.com/Just-Fun-Story-Accidental-Revolutiona...

However there is also Coherent Unix. I remember adverts for it as the full back page of so many magazines of the day, for $99. I always was on the fence about buying it. One of the big drawbacks of Coherent was that it didn't have any networking.

Then, Linux came out, and it was shockingly shortly after that that Linux surpassed Coherent, and Coherent vanished seemingly overnight.

https://en.wikipedia.org/wiki/Coherent_(operating_system)

"Coherent remains true to Unix's roots and eschews local area networking, graphical user interfaces, menus, mice, and many of the other amenities that present-day DOS users and Unix users have come to expect of modern software."

It goes on to describe that "learning Unix from Coherent would be a bumpy road" and "nor could you realistically use Coherent to automate a small business".

Coherent was closed-source, so you couldn't take an imperfect incomplete system like Linux and help improve it. Linux attracted contributors early on not because Linux at the time was so great, but because people liked to work on this kind of thing.

$99 is ~$230 today. I know software used to be a lot more expensive in general (e.g. $25 for PKZIP in the late 80s), but $230 for a bare-bones OS for playing around is only "cheap" relative to Enterprise Unix™® pricing.

[1]: https://books.google.ie/books?id=JZxkO0PpksUC&pg=PT61&redir_...

I was a customer and it was very good, I use to use kermit to dial into a Sun System at work and it would call me back so I would avoid charges. That was to get USENET and general Internet access. Fun times. It is too bad they had to fold, be to be honest by 1995 the writting was on the wall,

When they folded, I went to Slackware, which I still use. Back then, Linux was changing almost daily.

Coherent had one thing that Linux could not do, but via USENET I asked and someone supplied a patch for me within days. It would allow me to use 2 monitors with 2 cards on my 386sx, a BW and a VGA just like what Coherent was able to do.

People were far more accessible back then until the fortune 500 companies got involved in the 2000s.

Bless those contributors who were able to keep it up. They're the ones who should have Big Tech salaries, not those of us who gave up and sold our souls to Ad tech.

The first release of 386BSD was in 1992, so we are talking about a short time window of 1-2 years here. And of course FreeBSD and NetBSD forked from it shortly after that.

[1] https://en.wikipedia.org/wiki/386BSD

[2] https://www.antman.se/abc-klubben/linux/linuxdoc/linus.html

[1] https://en.wikipedia.org/wiki/Interactive_Systems_Corporatio...

pretty good documentation in a big doorstop they called "the lexicon.", a decent c tool chain, a custom port of emacs (microemacs they called it).

only other unix i played with before that was some port on the atari st, which incidentally was a first class platform for the mark williams c compiler (mwc was the company behind coherent).

it fit on two disks. where i had a massive stack of floppies for slackware (kernel 0.99pl2). linux was more fun- it was open source by and of the internet.

fun times!

And follow they did! The Great Linux Distribution Timeline at <https://upload.wikimedia.org/wikipedia/commons/1/1b/Linux_Di...> presents an impressive graph of the countless distributions that descended from Slackware, Debian, etc. This graph is still actively maintained by the community here: <https://github.com/FabioLolix/LinuxTimeline>.

The early days of Linux - https://news.ycombinator.com/item?id=35557848 - April 2023 (224 comments)

when 386BSD got it's act together, it became less apparent but very quickly impetus got behind the distro "scene" and Linux took off like a rocket.

Those first early days, no real network stack, minimal desktop support, it was line-and-ball. Mainly, I think the BSD lawsuit/licence issues was the negative energy source in BSD occupying the ground.

I had an IBM PC/RT at the time running the 4.3bsd port which was just lovely. But such clunky hardware. PCs were coming into their own and IBM didn't want to keep playing in multiple OS and had AIX coming through and the ROMP chip was getting old. Once graphics display cards on PCI emerged along with decent cheap ethernet cards, and then ethernet on the motherboard I think it was game over. (although for obvious reasons most people's experience of network was modem on serial)

A relevant key point regarding why Linux actually took off, versus other UNIX clones.

And how that eventually came around with Red-Hat's acquision and its influence on GCC, GNOME, Wayland, RPM,... contributions.

  Discussed at the time (of the article):

  The early days of Linux - https://news.ycombinator.com/item?id=35557848 -   
  April 2023 (224 comments)

Yet some people await for a second coming.

I find it interesting how academic circles back then were really not betting on the longevity of x86.

Maybe had BSD not be fighting AT&T, or Microsoft bothered with providing a proper POSIX experience in Windows NT insteaf of making the minimum for goverment contracts, and Linux history would have been much different, including its impact on cheap UNIX servers.

The joy of startx command working (sometimes)

Fun times.

I chose slackware like many french speaking fellas around. Oh, the times.

PS: debian all the way since.

Now I'm happy with Debian.

"I was there, Gandalf. I was there 3000 years ago" XD

The Linux desktop has had a rough history. The original problem was the original KDE using a non-FOSS Qt, which led to the development of GNOME, splintering the FOSS desktop ecosystem into two competing desktops, spreading already-thin resources. Granted, things were good in the mid-2000s, when GNOME 2 and KDE 3 were current, and when it appeared that the Linux desktop would be poised to make significant inroads against Windows XP, just like how Firefox in 2005 formed a formidable challenge to Internet Explorer. Even OpenOffice appeared to seriously challenge Microsoft Office at the time.

In my opinion, there have been many setbacks since the late 2000s. GNOME 3’s very controversial changes led to the splintering of the GNOME/Gtk community (Cinnamon and MATE). KDE 4 was not well received, though KDE 5 is widely praised. GNOME and other projects (though not KDE) was heavily influenced by mobile computing paradigms led by a misguided belief that desktops will become obsolete. The transition from X11 to Wayland has been quite rocky, to say the least.

Interestingly enough, what has improved the competitiveness of the Linux desktop these days has been the shift away from native apps and toward both the Web and Electron apps. While as a long-time former Mac user I lament the decline of native apps and the loss of consistency, apps built on Web technologies are cross-platform or even platform-agnostic. If one doesn’t need native Windows and macOS apps, then the barrier to switching to a Linux desktop is much lower. Even the gaming situation has dramatically improved on Linux.

Without Steam's built in Proton compatibility layer no gamer would realistically use Linux.

Outside of work I'm probably 70% Linux and 30% Windows. Music production is much better on Windows, but software development is miles easier on Linux. Messing around with Ubuntu during my teenage years was a key part of my later career!

------------ Please ignore just jotting down for future myself ---------------

(Actually I want to study Win NT more than Linux as it brought me good memories, but Win NT was designed as a mature product from the beginning, and even the leaked 3.5 source code is way too much for me.)

I checked the version history and 0.11 seems to be a good start. Reasons:

- There is this book (https://download.oldlinux.org/ECLK-5.0-WithCover.pdf) which follows the style of Lion's Unix v6 book and uses the 0.12 kernel as reference;

- 0.11 doesn't have a virtual memory or a virtual terminal, and they are among the software I'd like to implement;

- 0.11 does have the basic stuffs so I assume it is not too difficult to set it up with QEMU and QEMU-gdb;

The list of software to implement in mind:

- a Debugger (I'll probably use QEMU-gdb, but if I can write a simple step-in debugger maybe it makes it a bit easier to dogfood? Still too early though as there is no compiler on it)

- a Virtual Memory system (following the MIT xv6 lab, maybe a bit more advanced as I will complete the MIT xv6 labs before pursuing this venture);

- a Virtual terminal (not sure how much I can implement, maybe just a basic one with limited control sequence support);

- a USB driver (not sure if it's possible for 0.11, I have never written a driver before);

- a mouse driver (again probably too difficult for me);

- a LOT of CLI programs as warm-up. Basically simpler versions of modern ones (less switches);

- a text editor. I never liked the key layout of emacs or vim so this is going to be my own thing. I care a lot about moving caret around without using a mouse;

- a very basic GUI system. Nothing fancy, just something that can show windows and some widgets. Maybe don't even care about multiple windows. Showing one window with a menu and some buttons probably needs a lot of effort;

- a compiler/linker of a simple language (can't dogfood without an editor, a compiler, a linker and a debugger I think)

Preparations:

- I need to complete all MIT RISC xv6 labs;

- I need to figure out how to write drivers for USB and mice;

- I want to investigate how X11 and Win32 GUI were implemented and figure out the architecture;

- I need to figure out how to write terminals and text editors. There are tons of tutorials online so shouldn't be too tough;

- I need to study compiler theory, just continue reading Crafting Interpreters should be good enough;

- I need to figure out a comfortable development and debugging environment;

https://www.amazon.com/Developing-32-Bit-Operating-System-Cd...

Can probably find it on a shadow library. Getting something to boot on x86 is a little tricky, because IIRC the CPU comes up in real mode, so you have to create your own boot sector and bootloader before you can even execute your kernel.

That book + Linux inspired me to write my own OS in 1995. My goal was to have an OS that booted directly into a modern GUI and fitted (with basic tools) onto a single 1.4Mb floppy, which I did manage.

Now that I think more about the project, probably the first obstacle is how do I compile the thing...someone must have tried that, or the book must have a section about that. But I'll study xv6 first, just to give me some practices.

Windows NT Design Workbook:

https://computernewb.com/~lily/files/Documents/NTDesignWorkb...

Tale of Two Kernels:

https://youtu.be/HdV9QuvgS_w

And I'm sure you've picked up a copy of Windows Internals. Right now I'm waiting for a copy of Inside Windows NT which is the first edition of that series as well as Inside the Windows NT File System, both by Helen Custer.

Linus Torvalds had wanted to call his invention Freax, a portmanteau of "free", "freak", and "x" (as an allusion to Unix). During the start of his work on the system, he stored the files under the name "Freax" for about half of a year. Torvalds had already considered the name "Linux", but initially dismissed it as too egotistical. [2]

In order to facilitate development, the files were uploaded to the FTP server (ftp.funet.fi) of FUNET in September 1991. Ari Lemmke at Helsinki University of Technology (HUT), who was one of the volunteer administrators for the FTP server at the time, did not think that "Freax" was a good name. Therefore, he named the project "Linux" on the server without consulting Torvalds. [2] Later, however, Torvalds consented to "Linux".

To demonstrate how the word "Linux" should be pronounced ([ˈliːnɵks]), Torvalds included an audio guide with the kernel source code. [3]

[1] https://en.wikipedia.org/wiki/History_of_Linux#Naming

[2] Torvalds, Linus; Diamond, David (2001). Just For Fun - The Story Of An Accidental Revolutionary. New York: HarperBusiness. p. 84. ISBN 0-06-662072-4.

[3] Torvalds, Linus (March 1994). "Index of /pub/linux/kernel/SillySounds" (https://www.kernel.org/pub/linux/kernel/SillySounds/). Archived (https://web.archive.org/web/20091008074754/http://www.kernel...) from the original on October 8, 2009. Retrieved August 3, 2009.

Every field office had 2 or 3 PC's dedicated to the task of providing a "remote terminal" to our development team, so that we could dial in remotely and do administrative things to the Progress 4GL database server that was running on another dedicated system for the task, usually a fat Novell Netware server with 4 or 8 "huge" disks of 640Megabytes dedicated to the databases.

It was my job to monitor the rows and rows of PC's on our side of things, that were used for remote admin.

Honestly, I despised this job.

I'd spent most of the 80's learning C and Unix programming - my first Unix machine was a MIPS Computing Magnum pizzabox running risc/OS. $15,000 worth of machine, sitting on my desk - but yet I still had to walk down the corridor and hack around on cheap DOS PC's just to get things fixed across the country in some hellish New Hampshire closet.

At least it paid me well.

Well enough that I could go to the regular PC Computer swap meets out in Pasadena every weekend or so, and buy myself the parts needed to build a "decent computer for home" - a 486 with 8Megabytes of RAM, a few hundred megabytes of disk space. A CD-ROM drive, maybe a high-end video card and monitor capable of 1025x768 and 256 colours. A modem, so I could work from home, run my own DesqView instance, have my own multiple-terminal connections to all those troublesome field offices.

I'd also make regular (weekly) jaunts to The Opamp Bookstore, in Hollywood - one of my favourite places in all the world. It was there I'd wandered into the "Unix" section and seen some magazines and books on the MINIX operating system, and it was there that I learned about the rise of Usenet as a place to get to meet other hackers, C programmers, Unix nerds. It was also where I'd get my 2600 Magazine fix, monthly.

One day, I got my home PC finally built, booted, and running. I connected to my mailserver, wnsnews.com .. checked my mail. A fellow operator had sent me a message "Check out the minix-list, they're talking about a new unix-like kernel you might be interested in..."

So I subscribe, and lo and behold, there it was - the link to ftp.funet.fi, Linux in its very early stages. It took me about a day to download it, about a week to figure out all the various floppy disks I'd have to make to get the drivers for everything set up, and a whole weekend of faffing around to get things installed alongside my Desqview setup, on a brand new hard drive.

I was hooked. It didn't have everything - X11 wasn't ported yet - but it had just enough for me to port some of my own C code, see that things could actually work like my MIPS pizzabox, albeit on my cheaper 486, and .. maybe .. someday .. someone would port "Remote Control" or even the Desqview terminals, to this brand new 'Linux' thing.

Some months later, in the middle of being shut down entirely by my fellow grey-beard programmers, who wanted me to stop fiddling about with all the spare parts in our "Remote Control" room, I learned about this brand new Linux 'distribution', Yggdrasil.

A trip to OpAmp book shop, and there it was: the bootable floppy and the full Linux filesystem, included on CD, to turn any of our machines into a working Linux system - with X11 onboard! And, probably working, if I chose our SVGA card settings right! What a revelation! I could use my home 486 as a decent terminal for my pizzabox! I could use telnet to get into the Novell servers across the other side of the country! I could endlessly annoy my grey-beard associates with the idea that "Linux is the Future, man!", over and over again, until they finally lock me out of the Novell server and decide I was only good for Unix things, as punishment, because as they knew full well "Unix is Dead, man!"

Well, the decades have rolled by, and what a death it is. Linux is everywhere. I have used it every day since that first post on the minix-list I noticed, the one about funet.fi. I've made a living with it, and I've done amazing things, boring things, mundane things, and normal things. "Netware is Dead, Man! Desqview is Dead, Man! DOS is long gone, man!"

Well, alright, MIPS Computing didn't last long, and that pizzabox was well and truly (sadly) de-commissioned. But, with the profits I made in SoCal during the 90's, putting Linux to work from my own little consulting office in One Wilshire, I bought all the SGI toys I could want .. and I still have them.

Curious, though, that I mostly use Apples' Unix as a front-end to access them all, heheh ..