Linus Torvalds and friends tell The Reg how Linux solo act became a global jam session

If you know anything about Linux's history, you'll remember it all started with Linus Torvalds posting to the Minix Usenet group on August 25, 1991, that he was working on "a (free) operating system (just a hobby, won't be big and professional like gnu) for 386(486) AT clones." We know that the "hobby" operating system today is …

[DJV]

That cracked me up!

"'Well, I'm the other presenter, and these are the people that were in my audience, but we all would rather hear Linus's talk.'"

I would have loved to have been there to see that!

[phuzz]

I was only a kid then, and mostly concentrating on my Amiga, but I remember that there was all sorts of 'bedroom experiment' software and OS's, and to this day I don't know why Linus's experimental OS became the Linux we know today. At that point it wasn't even clear that x86 would have such staying power, RISC was talked about as being more future proof.

The future could have gone in many different directions at that point.

[Doctor Syntax]

"and to this day I don't know why Linus's experimental OS became the Linux we know today"

I think that is largely down to Linus himself and they way he was prepared to make it a communal effort.

Another factor might have been Minix which AFAICR was not so much an academic OS as an educational one. I wonder how many of the early contributors were studying that.

A final aspect must have been that they were familiar with Unix but even the SCO version was expensive. If SCO had grasped the fact that hey could have sold it at scale at a desktop OS price (there was a GUI desktop available) they could have cleaned up.

[jake]

"A final aspect must have been that they were familiar with Unix but even the SCO version was expensive."

Mark Williams Company's Coherent wasn't expensive at $99/seat, so that wasn't it entirely.

I rather suspect a large portion of it was that it worked with the already well-known GNU tools, coupled with the fact that it ran on cheap or free, used, bound for the scrap heap 386 PCs, and unlike the early x86 BSDs (NET/1, NET/2, BSD386 and 386BSD) it had no perceived potential for Ma Bell's lawyers to come sniffing around. Basically, it ticked all the boxes for both starving students AND startup companies in the still young SillyConValley.

[Gene Cash]

I don't think Coherent actually got any air time or buzz marketing wise, so few people knew about it. I would certainly have been interested in it, but I didn't hear of it until the company was out of business.

[jake]

Coherent had very prominent full-page adverts on the back cover of Byte magazine, and sometimes on the inside front cover. It was mentioned many times in both DrDobbs and SysAdmin magazines, both of which also ran ads for it. It had it's own Usenet group, and a couple of IRC channels. It was very well known in the *nix field[0].

What it didn't have, until it was too late, was a TCP/IP stack. It made do with serial communications, and came with Taylor UUCP.

_{[0] Not that the field had all that many folks in it at the time ...}

[vulture65537]

I remember SysAdmin about 1997 carrying Red Hat ads saying this is the OS for people who write their own device drivers faster than the vendor can send them.

[Joe Gurman]

Maybe that….

….was when you became aware of Silicon Valley as an incubator of tech startups with the assistance of venture capital, but that really started with the HP garage and Stanford engineering dean Frederick Terman’s recruiting of VC in 1951.

By tech outfit standards, it was pretty long in the tooth by the mid-1990s.

[jake]

Re: Maybe that….

I grew up in Palo Alto. I remember the cherry orchards, the canning industry, and apricots drying in the sun.

Trust me, Silly Con Valley was still in the prototype stage in the early 1990s. Beginning to take off, yes, but just barely fledged. Consider that analog telephone central offices were still very common, the screach of modems connecting was normal, floppy disks were still in use by almost everybody. and Intel 286 computers were still being replaced by 386s at businesses.

[david 12]

Mark Williams Company's Coherent did not have virtual memory It was limited to the amount of memory you (could afford to) put in your PC.

As such, it was a toy unix for people who needed or wanted a pc unix to complement the unix they used at work or at university, but it didn't offer a compelling alternative to DOS or Apple.

[jake]

The last couple of versions of Coherent did indeed have virtual memory, but they hadn't managed to implement demand paging yet.

[xasperated]

Coherent didn't have Networking. So that limited is usefulness. I used it for a year or so but moved to minus pretty quickly

[steelpillow]

In his autobiography Just for Fun, Linus said how he tried MINIX and wanted to make some improvements, but Andrew Tannenbaum would not let him.* This was what pushed him into starting from scratch. He and RMS (creator of GNU and the GPL) were kindred spirits in their disgust at being burned by proprietary lock-out, and their determination to avoid it themselves.

*AIUI Andrew did eventually release MINIX under an open license.

[ICL1900-G3]

Minix

I used Minix in the early nineties on my Atari Mega ST. It was remarkably capable and there was a surprising amount of software for it. I had Tanenbaum's book which guided you through how it all worked. I found it fascinating and am grateful to all concerned.

[mtp]

Minux

I used Minix on my Atari ST for many years and had much fun fiddling with it then in the early 90s I installed Slackware on a 386 PC from a big stack of floppies and Linux has been my main OS ever since. I even used it at work for many years.

[werdsmith]

I was only a kid then, and mostly concentrating on my Amiga, but I remember that there was all sorts of 'bedroom experiment' software and OS's

I was introduced by a BBS admin friend who copied me a set of Slackware 3.5 inch floppies. I booted to a command line on one disk I think, my memory is hazy but I do recall seeing the file names in different colours when I did ls.

[Eric 9001]

I know

Linus's experimental kernel became the kernel we know today because of GNU - once combined with GNU, the resulting GNU/Linux OS was quite good, but lacking due to deficiencies in Linux, which attracted thousands of developers to work on Linux.

>At that point it wasn't even clear that x86 would have such staying power, RISC was talked about as being more future proof.

GNU was carefully planned to be portable between machines, to prevent a repeat of the past disaster of all the work on an old free OS, written in assembler, turning to dust as soon as the specific machine it was written for went out of support.

Linux meanwhile used an ancient monolithic design - as it was quick and dirty.

x86 is dead (although it hangs around as a corpse as cruft) - AMD64 is a vastly different architecture - most of the instructions even end up getting executed as RISC micro-ops.

>The future could have gone in many different directions at that point.

There is no way to predict how the future would have gone - maybe GNU would have been successfully destroyed, but without Linux, I reckon the GNU developers would have likely allocated enough development effort to finish Hurd, to result in an GNU OS pretty similar to GNU/Linux today, except much more free.

[Throatwarbler Mangrove]

Re: I know

Okay, I'll bite: how is Linux not "free"?

[Eric 9001]

Re: I know

Linux is not free software, as it contains proprietary software without source code, disguised as arrays of numbers. Just a few cases are;

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/powerpc/platforms/8xx/micropatch.c

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/usb/dvb-usb/af9005-script.h

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/common/b2c2/flexcop-fe-tuner.c#n227

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/i2c/vgxy61.c#n115

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n122

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n12

Also, if you look at a lot of Linux drivers, you'll realize that half the driver is missing and the other half is distributed in; "linux-firmware.git", which is a massive collection of proprietary derivative works of Linux without source code, under a proprietary license, that can't be anything but part of Linux (there's also a handful of free software programs under GPLv2-compatible licenses in that repo too).

If you want Linux with freedom, you'll need; https://gnu.org/software/linux-libre

[Throatwarbler Mangrove]

Re: I know

As I understand the model, one has the option of not using proprietary drivers. Certainly, when I've installed various flavors of Linux recently, I've been offered the choice of whether to do so. Therefore, it seems like one has the choice of installing either ideologically pure Linux, using only open drivers, or installing the closed source drivers and suffering the foul taint of corruption. I'll admit that I might be mistaken, however.

Anyone more knowledgeable about such things want to weigh in?

[Eric 9001]

Re: I know

I'm clearly knowledgeable about such things, why not ask me?

Those were proprietary GNU/Linux distros, mostly only differing by what DE and what package manager or whatever is installed by default (you can install whatever DE and whatever package manager you want on any GNU distro, although the default configuration of some distro's may make installing certain things very painful).

Most proprietary distros use the exact same Linux, only differing by versions and maybe with some trivial patches.

The user is not given a choice - there's proprietary software in the Linux tree (as linked) and in therefore, the typical generic kernel used for most distro's has such proprietary software compiled in and has all the proprietary software loading machinery enabled too.

Most proprietary distro's, like Debian, don't even give the user the choice whether to use proprietary drivers - as the only available official installer comes with most or all of the proprietary software and if the installer detects if the hardware could run a certain kind of proprietary software, it automatically permanently installs it and loads it *without asking*.

The "choice" given in such installers is typically whether to install even more proprietary software of a slightly different kind - for example the proprietary nvidia driver (if you click no and there's a Nvidia GPU, I figure it uses Noveau with proprietary software - even though Noveau seems to operate without proprietary software even for newer GPUs).

Please advise which distro and I'll find out for you what exactly that option in the installer does.

If you want a free Linux, that gives the user a choice, the only version of Linux that does that is GNU Linux-libre - as it doesn't contain any proprietary software or proprietary software loading machinery - although that needs to be combined with a free GNU distro that doesn't ship other proprietary software by default to give choice.

With GNU Linux-libre, the user is given the choice whether they want to load proprietary drivers - although most choose to not do so.

[Throatwarbler Mangrove]

Re: I know

You're also clearly a crank with an ideological axe to grind who came in here making unsubstantiated, bald-faced assertions as well as broad ad hominems against Linus Torvalds personally, so you'll forgive me if I don't take your assessment as gospel.

[Eric 9001]

Re: I know

My assertions are substantiated - I even linked proof. If I need to spell out a specific assertion further for you, please quote it.

I haven't even got GNU/Axe out yet.

Obviously I would link to the specific source files - so you wouldn't even need to take my word - as you could check yourself.

I'm sorry, but detailing why I don't glaze Linus, by pointing out all the bad things he has done as well as the few good things, is not "ad hominem".

Meanwhile, referring to me as a "crank" and dismissing my arguments over character and not substance, is the purest kind of "Ad hominem".

[Antron Argaiv]

Re: I know

Linux is free enough for me.

I remember downloading diskette images from ftp.tsx-11.mit.edu It was great when Walnut Creek CD started selling CDs of Linux for like $10. Yeah, I'm old. Started playing with Linux around 1993, and it's been my daily driver for years.

Thanks, Linus and friends.

[xasperated]

Re: I know

I still have one of the first walnut creek cds

[3arn0wl]

It's a remarkable thing

Thank you Linus, ard everyone else who has made a contribution over the years.

[Eric 9001]

Re: It's a remarkable thing

Linus really scrapes by when it comes to deserving thanks, as for every good contribution he has made, he has made at least ten diminutions.

Thank you Linus for deciding to relicense Linux under GPLv2-ambigious in 1992 and your later endless code reviews (but no thank you Linus for allowing Linux to become proprietary again in 1996 and changing the license to GPLv2-only in 2000 and ensuring that the -only part is only even enforced against freedom and not against proprietary software).

Thank you Linus for releasing git as free software, even though it really would be much more useful had it been licensed GPLv2-or-later, but GPLv2-only is a free license after all (but remember that git was to clean up a mess Linus caused - as before git, he encouraged the proprietary software and SaaSS Bitkeeper for Linux development, which predictably had consequences).

Thank you Linus for work on that Suba diving program.

I can't think of anything else to thank him for - anyone else have any other suggestions?

Thank you GNU developers for your long service to freedom, delivering free software without fail, no matter what.

[LionelB]

Re: It's a remarkable thing

> I can't think of anything else to thank him for - anyone else have any other suggestions?

Um, how about for creating Linux and overseeing its development and maintenance over the past 35+ years?

[Eric 9001]

Re: It's a remarkable thing

Him originally releasing Linux as proprietary software was not deserving of thanks (as a developer that releases proprietary software deserve punishment for releasing a harmful program, not thanks) - he only deserves thanks for re-licensing it to a free license.

If you read my post, you'll realize that I did thank him for all his code reviews (what overseeing development and maintenance is).

So no-one can make other suggestions huh?

[LionelB]

Re: It's a remarkable thing

I suggest that you have a massive chip on your shoulder.

[werdsmith]

Re: It's a remarkable thing

Someone seems determined to deny Linus his due credit.

[Antron Argaiv]

Re: It's a remarkable thing

Sounds like sour grapes to me.

Nobody's ignoring the contributions of Gnu...what do you think the "g" in gcc stands for? Linux is a team effort, always has been.

[Eric 9001]

Re: It's a remarkable thing

The contributions of GNU and not being ignored - rather those are explicitly excluded, just like what you did.

GNU/Linux or LiGNUx or just GNU has been a team effort, but for some reason the GNU is always excluded in favor for only the name of only one of the kernel's GNU supports (despite the confusion that causes and the common factual inaccuracy resulting).

[Anonymous Coward]

Glorious article thank you

A wonderful glimpse into a fascinating time.

[Bran Muffin]

A great example

Although I am not a Linux user, I can still appreciate Linux (and Linus, too), for being what it is today. Linux should serve as an excellent example to every aspiring computer scientist, user, developer, etc., of what can be done with dedication and hard work. Expecting AI to generate the "next big thing" is a trap waiting for the next new student--computer systems with drive, ambition, and actual intelligence are a long, long way off, if such a thing is even possible.

[Eric 9001]

Re: A great example

It's only an excellent example as to what you level of credit you can achieve with careful selection of the right buzzword, a few years of hard work (that mostly ends after a few years, but don't worry, the fat salary will still remain) and luck - while other people do almost all of the work for you - you can claim all the credit for yourself with the right buzzword.

It was GNU that had the real dedication since 1983 and did the really hard work (and it is still fighting hard despite the continued attempts to destroy it) - but of course it's either cut out of the credit completely, or is dismissed as much as possible.

[m4r35n357]

Re: A great example

Are you getting paid to discredit GNU?

[Eric 9001]

Re: A great example

I'm not getting paid.

I don't see how pointing out what GNU did discredits it.

Is the problem me not meeting a corporate-approved tone?

[alisonken1]

Re: A great example

Interetsing.

GNU was working on the Hurd kernel. How far did they get?

Linus used GNU toolchain and libs and gave GNU a working kernel.

And last I checked, Linus did not make the kernel proprietary - it's always been free to download/use/modify without cost. The only difference was switching from his own personal license to the easier-to-enforce GPL (mainly because of the foundation that keeps the GPL valid).

As for "corporate-approved tone" - did you have a specific one in mind? Like Slackware or Debian?

Your earlier post:

"Him originally releasing Linux as proprietary software was not deserving of thanks (as a developer that releases proprietary software deserve punishment for releasing a harmful program, not thanks) - he only deserves thanks for re-licensing it to a free license.

If you read my post, you'll realize that I did thank him for all his code reviews (what overseeing development and maintenance is).

So no-one can make other suggestions huh?"

Can you defend your position on "releasing Linux as proprietary software ..."?

Again, Linus did not make the kernel proprietary at any time. Talk about someone talking out of both sides ...

[Anonymous Coward]

Re: A great example

Please do not feed the troll.

It can't be reasoned with, because that is not it's goal.

All it wants to be is noticed. Responding to it feeds that need. The more you respond, the more it posts.

And no, the rest of us don't need you to let us know it's full of shit. We all already know that. That's why we don't respond to it.

TINW

[jake]

Re: A great example

Oops. One of the cats must have clicked the "Anonymous" button.

ElReg won't allow me to change it back, even within the edit window.

I could withdraw it and then repost, but instead I'll just put my stamp on it this way.

[Eric 9001]

Re: A great example

I'm not even trolling. If I was trolling, I would go with "Anonymous".

My goal is to inform people that Linux is proprietary software and the OS is GNU (as most are too cowardly to tell them), no matter how much such reality is denied and disliked.

All my posts contain reason - rather you're the one who rather cannot be reasoned with, as you have been presented with the facts, but you deny them and *I'm* full of shit?

[Eric 9001]

Re: A great example

Hurd development continues - 64 bit is now working and SMP shall soon be ready, but don't forget that GNU also develops other kernels, such as GNU Linux-libre and the kernel of the GRUB OS.

Hurd for the moment isn't needed, thus it is little developed, but I reckon it's on track to be ready by the time GNU Linux-libre needs to be replaced.

>Linus used GNU toolchain and libs and gave GNU a working kernel.

He also used the rest of the GNU software and even then, his kernel wasn't useful at all until GNU developers helped him out and ported every last GNU package to work with Linux.

>And last I checked, Linus did not make the kernel proprietary - it's always been free to download/use/modify without cost. The only difference was switching from his own personal license to the easier-to-enforce GPL

>Can you defend your position on "releasing Linux as proprietary software ..."?

Free means freedom - only being concerned about things being gratis is very shallow.

You didn't check very well; https://www.kernel.org/pub/linux/kernel/Historic/old-versions/RELNOTES-0.01

This kernel is (C) 1991 Linus Torvalds, but all or part of it may be redistributed provided you do the following:

- Full source must be available (and free), if not with the distribution then at least on asking for it.

- Copyright notices must be intact. (In fact, if you distribute only parts of it you may have to add copyrights, as there aren't (C)'s in all files.) Small partial excerpts may be copied without bothering with copyrights.

- You may not distibute this for a fee, not even "handling" costs.

It could be argued that the license wasn't even valid, as "small partial excerpts" may still fall under copyright (thus meaning that the license encourages copyright infringement) and it contains a misspelling. If a license isn't valid, then there is no license; https://www.gnu.org/licenses/license-list.en.html#NoLicense

The license violates freedoms 2 & 3; https://www.gnu.org/philosophy/free-sw.html#four-freedoms as it restricts legitimate forms of distribution.

Selling copies of source code under a free license and changing fees for handling costs etc for software with complete corresponding source code available under a free license are legitimate activities and if you can't carry out those legitimate activities, it is not free software and it therefore it is proprietary software.

>The only difference was switching from his own personal license to the easier-to-enforce GPL (mainly because of the foundation that keeps the GPL valid).

Yes, in 1992 he relicensed from a proprietary license to the GPLv2-ambigious, making Linux free software, but Linux didn't remain free software for very long.

In 1996, the first of many proprietary programs without a license (https://www.gnu.org/licenses/license-list.en.html#NoLicense) and without source code were added to Linux;

`wget https://mirrors.edge.kernel.org/pub/linux/kernel/v2.0/linux-2.0.tar.xz ; tar xf linux-2.0.tar.xz ; nano linux/drivers/net/dgrs_firmware.c` - first proprietary program in Linux, dated;

"04/15/96 01:59:21"

https://web.archive.org/web/20060126231558if_/http://wiki.debian.org:80/KernelFirmwareLicensing

cpu-rec.py;

dgrs_code full(0x1b5c0) MIPSel chunk(0x15000;84) MIPSel

If any part of a kernel is proprietary software, the whole thing is then rendered proprietary, as the users cannot exercise the 4 freedoms with all of the kernel.

>Again, Linus did not make the kernel proprietary at any time.

If Linux is free software, then you'll be able to find the source code, under a GPLv2-compatible license, of the following programs disguised as arrays of numbers;

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/powerpc/platforms/8xx/micropatch.c

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/usb/dvb-usb/af9005-script.h

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/common/b2c2/flexcop-fe-tuner.c#n227

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/i2c/vgxy61.c#n115

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n122

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n12

Also, if Linux is free software, you'll also be able to find the source code, under a GPLv2-compatible license, of all the proprietary derivative works of Linux here; https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/

But you can't, as Linux is proprietary software.

[David Hicklin]

Re: A great example

> of what can be done with dedication and hard work

Of course things were a lot simpler back then and we did not have the Mega IT corps (they were still little and growing up back then) to dominate the landscape.

8 bit home micro's, even PC's were 286/386 going 486 at a pace you could keep up with and understand how they worked at the hardware level.

[okand]

I was born in 1992 so this all happened before I was able to pay attention to any of it; I love hearing about these days though. In like 2007 or so I had to read and write a book report and I picked the Swedish translation of "Just For Fun".

[gosand]

This article made my day. It's good to hear origin stories about my favorite operating system and my peers*.

*in age only... I am 2 weeks older than Linus.

[mickaroo]

Hardcore Penguinista

I messed with Linux back in the late 1990’s, but it was SO MUCH DIFFERENT from what I was used to (DR-DOS) that I didn’t commit. Then in the early 2000’s, necessity forced me to embrace Unix (Sun Solaris). After that, Linux was a walk in the park.

SUSE, Caldera, Corel, PC-LinuxOS, Ubuntu (various flavours), Mandrake, Debian (briefly), and finally… Linux Mint Mate.

I also got pretty good at building NDISWrapper from source and hacking Windows .inf files to enable various USB wireless dongles.

[Rory B Bellows]

Re: Hardcore Penguinista

I'd all but forgotten about NDISWrapper... it was the only way I could get the wireless on my old Acer Aspire laptop to work under Linux. Good times!

[David 132]

Re: Hardcore Penguinista

It's still tangentially used to get certain wifi cards to work under MacOS on Hackintosh builds, although of course those are a dying breed now as x86 fades from the Apple stack.

[captain veg]

Re: Hardcore Penguinista

Bob, I'd forgotten .inf files.

Unable to magic up the secret incantation to make my OEM CRT monitor work with any kind of Linux distribution that I tried I resorted to asking the manufacturer (or, at least, their European distributor). I actually asked, if I recall correctly, for the necessary parameters to get my Xorg config working. They proudly sent me a "device driver" file, which turned out to be a wholly useless Windows .inf.

I ended up buying a new monitor. My first flat panel.

-A.

[frankvw]

Re: Hardcore Penguinista

In the late 1980s I cut my teeth on a very nasty port of SVR4 that ran on an industrial Motorola VME bus development chassis that was never really intended for multi-user use but, thanks to plugin cards with a dozen or so RS232 ports, was pressed into service as such nonetheless because of limited budgets in Dutch education at the time (a situation which continued to this very day). I still remember the excitement when I discovered the elegance of the Unix approach: how everything is a file and how pipes could string a dozen commands together. Even on those cantankerous systems I was hooked on Unix. I set up an "undocumented" dial-in link so that I could access the Unix box from the cozy comforts of my bedroom. (This was in the heydays of BBS'es, one of which I ran as well, and serial modems ruled the roost.)

Then, through fellow students in Amsterdam, we got hold of a copy of Minix (on a lot of 5.25" floppies) which worked well as an example code base to teach OS development with, but often had stability problems at the time, especially with the 30MB RLL harddisks that had just begun to replace the ubiquitous 20MB ST-225. We also had a clunky IBM AT (a real one - it weighed a metric ton and a half) that ran Xenix and that everyone hated for some reason.

Then, not long after I got my first job, I visited a trade show in the US and came home with a copy of Slackware 0.18, and that was it. I just couldn't live with the limitations of DOS and Windows anymore. Two years later I was an IT contractor for clients with anything from NCR mini's running their own clunky SVR V ports to SunOS and Solaris boxen, and I've never looked back. I've worked on everything from AIX to HP-SUX, but Linux has always been there for me through it all.

Life with Linux is a lot easier these days. I remember compiling support for ISDN cards into the kernel, and configuring X by editing text files. But then as now there was always a large community willing to help you out if you really got stuck. And that is, IMO, why Linux is what it is today.

[Eric 9001]

Amazing, the article actually mentioned GNU

But of course, it contains the typical untruths and false statements.

>This clarified that anyone could study, modify, and redistribute the code as long as improvements remained free.

The re-licensing to the GPLv2-ambigious was not a clarification - Linux was originally explicitly licensed under an proprietary license, but he was getting nowhere, as GNU wasn't interested in supporting a new proprietary kernel.

After he re-licensed, GNU decided to port every last package to work with Linux out of goodwill (which of course Linus betrayed in 1996, by allowing Linux to become proprietary software again) and thus GNU/Linux was born.

>That made it possible for developers to build distributions

Even under the proprietary license, Linux could and was combined with GNU, which is what Linux's developers did in 1991 as part of development, but that didn't result in an OS anyone would want to use, as for example "getting Bash and GCC to run" wasn't possible until those were ported with a lot of work and even after that, the OS still sucked and wasn't worth distributing, as why distribute 2-4 broken GNU packages and a useless kernel?

People only bothered to start preparing and distributing GNU/Linux distributions, once GNU went and ported dozens of GNU packages and therefore bestowed the user with the proper GNU experience with all the GNU packages, including GNU Emacs - too bad most of such distributions went under the name of "Linux" - as buzzwords.

>that combined the Linux kernel with GNU tools

Well, there it is - the part where GNU is dismissed and downplayed - after all, tools are only a handful of GNU packages and those aren't even the exciting packages.

As mentioned about, what actually happened is that GNU was almost finished (it was just missing a kernel) and Linux was combined with GNU to make GNU/Linux (the limited amount of other free software that then existed, like Xorg (there really wasn't much of it ever since copyright was applied to software and thus made software automatically proprietary unless it was validly licensed under a free license, with the copyright information kept in order), was added later).

But, it is crucial that GNU must be dismissed and downplayed, therefore "the Linux kernel" is written to mean Linux and "Linux" is written to mean GNU (with or without Linux as the kernel).

[Androgynous Cupboard]

Re: Amazing, the article actually mentioned GNU

You may have noticed the title of this article? It's an interview with Linus and earliy Linux people, about Linux. We all know GNU is important. But it's not the topic of this story.

[Eric 9001]

Re: Amazing, the article actually mentioned GNU

The title doesn't really match the article - the article is not just about Linux development, it's about the GNU/Linux OS (referred to as "Linux"), except of course only Linus and early Linux people were interviewed, despite how GNU people are easier to reach for interview.

[NetMage]

Re: Amazing, the article actually mentioned GNU

Is there a way to block posters?

[Rory B Bellows]

Re: Amazing, the article actually mentioned GNU

Post no bills ?

[The Central Scrutinizer]

Re: Amazing, the article actually mentioned GNU

I wish. He is so annoying and full of it.

[Eric 9001]

Re: Amazing, the article actually mentioned GNU

No - it's likely scheduled far behind IPv6 support and being able to see replies without having to refresh the page and look.

Until then, you can look at the username and see that it's over 9000! and not read it.

[David Hicklin]

Re: Amazing, the article actually mentioned GNU

> Is there a way to block posters?

You can do that on the BBC comments now which is great, so have an upvote from me as it would be a good feature. Might make the threads a little disjointed .

[Dan 55]

Re: Amazing, the article actually mentioned GNU

The title doesn't really match the article - the article is not just about Linux development, it's about the GNU/Linux OS (referred to as "Linux")

That was this article from 2021 which includes a quote "making Linux GPL'd was definitely the best thing I ever did" from another interview from 1997.

I'm not sure what supposed to have happened in 1996, Linux was always GPL v2 except for the first year or so.

[Eric 9001]

Re: Amazing, the article actually mentioned GNU

In 1996 the first proprietary program of many was added to Linux and as that program did not come with source code, under a GPLv2-compatible license;

`wget https://mirrors.edge.kernel.org/pub/linux/kernel/v2.0/linux-2.0.tar.xz ; tar xf linux-2.0.tar.xz ; nano linux/drivers/net/dgrs_firmware.c` - first proprietary program in Linux, dated;

"04/15/96 01:59:21"

https://web.archive.org/web/20060126231558if_/http://wiki.debian.org:80/KernelFirmwareLicensing

cpu-rec.py;

dgrs_code full(0x1b5c0) MIPSel chunk(0x15000;84) MIPSel

Linux is now about half GPLv2-only and about half proprietary software, despite how the inclusion of *any* proprietary component violates the GPLv2 and automatically permanently terminates the license on distribution.

>making Linux GPL'd was definitely the best thing I ever did

He won't say it, but the main reason why it was the best thing he ever did, was because it convinced GNU developers to port every last GNU program to work with Linux and therefore made it possible to conveniently run GNU software and other free software, without being afflicted by the nasty proprietary license and also high prices of Unix's - as part of a GNU/Linux distribution.

In the first GNU/Linux distributions, a lot of things were broken or lacking due to deficiencies in Linux, which attracted thousands of developers to go and improve Linux.

[DoctorPaul]

Re: Amazing, the article actually mentioned GNU

You do realise that you're coming across as someone with a massive axe to grind don't you? Rhetorical question of course.

I don't owe an allegiance to any particular operating system, but what gave me a nice warm feeling about the article was how friendliness and cooperation between individuals could achieve so much.

And then you had to remind us of the dark side of the "community". Just wanted to ask - did you enjoy a good flame war back in the day?

[Antron Argaiv]

Re: Amazing, the article actually mentioned GNU

I hate to bring facts into this flamewar, but Linux made a design choice when they included proprietary drivers.

Yes, Linux to this day contains proprietary binary blobs.

The choice was: write network and display drivers from scratch, with minimal documentaion and no cooperation from the vendors (many of whome are in China), or make deals to include binary blobs and deliver an OS with up-to-date support for popular network and display devicest. The Linux developers held their noses and opted for the latter approach, even though they would have preferred to have all-open drivers.

It's just not reasonable to expect open drivers for highly proprietary chips where even "legitimate" developers can't get full chip documentation without licenses and NDAs. Many of these manufacturers write (or wrote) their documentation only for the market they were targetting: Windows PCs. Getting any other support from them was...unlikely. I know, I have tried. And often, the Windows drivers were...less thn perfect (I have heard them called "barely functional" and sometimes worse).

Personally, I think they made the right decision, as I remember the "good old days" when getting a network or display card up and running involved a whole lot of messing about, and you had very limited choices about which hardware would work.

[Eric 9001]

Re: Amazing, the article actually mentioned GNU

I'm bringing the facts, others are bringing flames, as they don't like the facts.

The Linux developers made the wrong decision, trading short term convenience for long term problems and inconvenience.

It takes a lot of work to reverse engineer hardware and write a free driver (or get a company to choose between coughing up the source code of their derivative work, or being unable to sell the product), but when it's done, everyone can enjoy the free driver and conveniently the network and display cards end up *just working without messing about*.

For example, ath5k & ath9k just work - while proprietary Wi-Fi drivers are all a buggy mess.

Had they instead from the start taken a policy that only free drivers were permitted and enforced the GPLv2 against companies that violated Linux's copyright (a company that sells routers needs Linux if they want the routers to be any good - while Linux doesn't need that company), more hardware would be supported and all of Linux's drivers would be free today.

[werdsmith]

Re: Amazing, the article actually mentioned GNU

OK Eric, you don't like Linus. We get it. Whatever your problem with him, it comes over very chip on shoulder. Get over it.

[dmesg]

Re: Amazing, the article actually mentioned GNU

>> [W]hich of course Linus betrayed in 1996, by allowing Linux to become proprietary software again ...

Um, last I looked, Linux was still under the GPLv2-with_nuance.[*] Are you perhaps a member of the Judean Popular Front for FOSS Purity?

>> [W]hat actually happened is that GNU was almost finished (it was just missing a kernel) ...

That word "just" is carrying a lot of weight there.

Look, I get it. The GNU project contributed a ton of essential early work, independent of Linux, yet "Linux" is the name (almost) everyone uses. It's not fair. "Lignux" was proposed at the time, but ... an OS pronounced as "lug nuts" or "lick nuts"? For a year or two I tried always saying or writing "GNU/Linux", but eventually realized I was swimming against the current. It's too awkward for conversation. Natural language evolution is going to smooth it into something easier.

A though to take consolation in: it's amazing what you can accomplish if you don't care who gets the credit. Apparently from Ralph Waldo Emerson but (fittingly!) usually attributed to Harry Truman.

[*] https://www.kernel.org/doc/html/latest/process/license-rules.html says:

The Linux Kernel is provided under the terms of the GNU General Public License version 2 only (GPL-2.0), as provided in LICENSES/preferred/GPL-2.0, with an explicit syscall exception described in LICENSES/exceptions/Linux-syscall-note, as described in the COPYING file.

This documentation file provides a description of how each source file should be annotated to make its license clear and unambiguous. It doesn’t replace the Kernel’s license.

The license described in the COPYING file applies to the kernel source as a whole, though individual source files can have a different license which is required to be compatible with the GPL-2.0:

GPL-1.0+ : GNU General Public License v1.0 or later

GPL-2.0+ : GNU General Public License v2.0 or later

LGPL-2.0 : GNU Library General Public License v2 only

LGPL-2.0+ : GNU Library General Public License v2 or later

LGPL-2.1 : GNU Lesser General Public License v2.1 only

LGPL-2.1+ : GNU Lesser General Public License v2.1 or later

Aside from that, individual files can be provided under a dual license, e.g. one of the compatible GPL variants and alternatively under a permissive license like BSD, MIT etc.

[Eric 9001]

Re: Amazing, the article actually mentioned GNU

>Um, last I looked, Linux was still under the GPLv2-with_nuance.

In 1996, Linus permitted the first proprietary software program without source code to be included in Linux, therefore making GPLv2-ambigious no longer the license.

Many more of such proprietary programs, without source code, disguised as arrays of numbers have been added since then, a few are;

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/powerpc/platforms/8xx/micropatch.c

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/usb/dvb-usb/af9005-script.h

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/common/b2c2/flexcop-fe-tuner.c#n227

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/i2c/vgxy61.c#n115

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n122

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/iio/proximity/aw96103.c#n12

Due to the proprietary software without source code contained, note that if you were to distribute Linux, your license of the GPLv2-only parts will therefore be automatically terminated permanently - if you want to distribute Linux without losing your license, the options are; https://gnu.org/software/linux-libre or maybe Debian's modified version.

There was a policy that proprietary derivative works are to be moved to the "linux-firmware.git" repo (note the selection of name - peripheral software for whatever kernel would be "firmware.git", while peripheral software that is part of Linux would be "linux-firmware.git"), but that policy still hasn't been implemented in over a decade.

>Are you perhaps a member of the Judean Popular Front for FOSS Purity?

I'm purer than "FOSS" degeneracy - free software!

>saying or writing "GNU/Linux", but eventually realized I was swimming against the current. It's too awkward for conversation. Natural language evolution is going to smooth it into something easier.

By saying and writing GNU slash Linux, I am swimming against a proprietary tsunami and what can I say but oh yeah!

The true evolution would be just GNU.

[Antron Argaiv]

Re: Amazing, the article actually mentioned GNU

JFC, Eric, give it a rest.

[Anonymous Coward]

GNU was almost finished (it was just missing a kernel)

That's like saying a car was almost finished - it was just missing an engine.

[Eric 9001]

Re: GNU was almost finished (it was just missing a kernel)

Linux is not the engine - when it comes to a car analogy, the kernel is closest to the transmission (as a kernel allocates the machine's resources via complicated logic, while a transmission allocates a machine's torque efficiently to the output shaft via complicated gearing) and the engine is rather closest to the CPU.

GNU was previously run on proprietary Unix's OS's, where each proprietary program had been replaced one piece at a time and almost everything was replaced, except for the kernel.

GNU/Car was therefore almost ready, it just needed a free transmission and Hurd development therefore begun - but Linus later appeared and wrote a quick and dirty kernel using a monolithic design, which became one of the transmissions once it was relicensed to the GPLv2-ambigious.

GNU/Car now supports several transmissions with more or less the same result and feel regardless of the transmission - but of course the whole car must be named after the transmission!

[storner]

Sending code to Linus

I started playing with Linux around 1994, and shortly after Windows '95 arrived with the VFAT long filename support. Someone wrote a new Linux filesystem driver for it, but didn't have the time to get it upstream and reworked so Linus would accept it - being bold and adventurous I took that job and after a bit of work put the patch in an email to Linus.

He quickly answered and asked for some modifications (improvements). I was completely star-struck that the famous Linus took time to direct me to how I could implement something better ...

The code did get merged after a couple of iterations. Later I did a lot of work on the kernel configuration tool to accomodate building modules, which earned me a listing in the CREDITS file - I consider that a high point in my software development career ;-)

(Icon: Closest one to a grey-beard)

[DS999]

Re: Sending code to Linus

I sent him a few fixes I found when I was helping out an engineer at the company I worked for at the time (would have been 1994/1995 timeframe) who was playing with Linux. I used a old free gnu fsf email I had signed up for a few years earlier and a fake name when submitting and they're still in the CREDITS file. I didn't use my real name/email because I was worried I'd get in trouble with my employer, who had one of those crappy contracts where they basically own everything you do (the first and only time I ever had to sign one of those)

The idea of "GPL" was still pretty new then, and I didn't want to find out what their lawyers would think about one of their employees "giving away" IP. Especially since it was technically done "on the clock"!

[steviebuk]

Missed it all

I was in college in the early 90s learning Pascal and Visual Basic on Windows 3.11, then eventually Windows 95. We saw the birth and growth of the Internet while there and the first between Netscape Navigator and IE. Looking back, it was good to be around then. But, sadly, I oddly missed the birth of Linux and never spotted any of that. I think it was because it was awhile before my parents could afford to buy us a PC for my course. And only having one and no Internet meant never found out about Linux. Even if had, wouldn't of had a spare PC to mess around with it anyway.

[David 132]

Re: Missed it all

Two words that cause those of us who went that route to involuntarily twitch, even to this day:

Trumpet Winsock.

[Antron Argaiv]

Re: Missed it all

There's a point to be made here:

Windows, for WAY too long, had no native TCP/IP stack, and required a third party memory manager (QEMM, anyone?). Linux had both, from early on.

Around 1994, my employer had me doing schematic capture on a 486, running Win3.1. Open too many pages in ViewDraw, and the OS would crash. I got tired of this, brought in a spare HDD I had at home, and installed (Slackware?) on it. Set it up for dual boot, and used it as a remote X display off our Solaris system (10BASE-T network), which also had Viewdraw, but was faster and did not crash. That was the moment I realised that Linux was far more useful to me in getting my work done, than Windows. Linux was free and worked better, even then, than the OS my employer had paid for.

[David Hicklin]

Re: Missed it all

> Windows, for WAY too long, had no native TCP/IP stack

Or the Novell IPX/SPX.. those were the days booting from DOS and using the Novell menu to launch Lotus 1-2-3 and Wordperfect 5.1 is how I cut my IT teeth (and holding down the day job!).

Later added HP-UX Unix and then IBM AIX before Suse at work and interest in Linux took off.

[jj_0]

Re: Missed it all

Well, actually you found out that Solaris was better than Windows

[DoctorPaul]

Re: Missed it all

Now that does take me back!

I put together the Windows 3 software stack that early UK ISP Nethead used to supply to their customers. Winsock, Netscape, Eudora, an FTP client and a menu/control panel to hook it all together.

Windows 3, so it couldn't have been my favourite IDE ever, Delphi, must have been Turbo Pascal for Windows.

[Anonymous Coward]

Re: Missed it all

"I put together the Windows 3 software stack that early UK ISP Nethead used to supply to their customers. Winsock, Netscape, Eudora, an FTP client and a menu/control panel to hook it all together."

Ditto, I wrote my first (and only?) VisualBasic program as the control panel (for WinSock, Mosaic (pre-Netscape), etc) for the very early UK ISP we launched in 1994.

[RegGuy1]

Slackware

That was an interesting article, thanks.

I remember Slackware and all those floppies. I think the first version I used was 0.85, but I didn't go for X11 because I think you needed 25 floppies. It was definitely a breath of fresh air, and I remember building a firewall with it, because you could define the simplest system to do just what you wanted, and not be bloated with stuff you don't need. Linux was perfect for stuff like that.

Beer for all those early pioneers. Ta very much for letting me have a platform to study and develop my skills that gave me a profitable career. Cheers!!

[Gene Cash]

FUNET

So I remember nic.funet.fi having a ton of stuff other than Linux, but for the life of me, I can't remember what, and Google gets more useless by the day.

Anybody else have info?

[captain veg]

Re: FUNET

> Google gets more useless by the day.

>

> Anybody else have info?

I believe that this has been analysed and characterised as "enshitiffication".

-A.

[David 132]

Re: FUNET

That FQDN rings a vague, dusty, long-forgotten bell with me, too.

Was it also an Aminet mirror, perhaps?

[jake]

Re: FUNET

Go take a look.

The obvious address is obvious.

WARNING! Rabbit hole!

[mjflory]

Auld Lang Syne

Thanks for the memories. In the early 90's I shared an apartment with a Comp. Sci. student who'd borrowed an AT&T 3B2 from the department and realized there was no way I could afford a Unix system. Then an old friend in the Unix division of Bell Labs told me his pals (you know, Dennis Richie and those guys) were excited about a new homebrew Unix clone, so before long I had a stack of SLS floppies on my desk and Linux on my fancy '386 PC. I spent weeks just getting an X Windows session running (just an empty screen, no applications) and realized that I could either get Linux running well or finish my post-doc, but not both. A few years later distros with graphical installers and other wonders appeared and I've been using them since.

[Chris Gray 1]

Different Path

1991 would have been too late for me (I'm *real* old!) I think it was the 70's that I would have been into all that OS stuff. My first computers were just for gaming, with cassette storage. I eventually got an Exidy Sorcerer system (CP/M) and a pair of 8-inch floppies. It's there that I started on my Draco compiler, etc. Then I spent a ton of money on an S-100 system to go with the dual 8-inch floppy box. Continued on with Draco, etc., and got a couple of plug-in graphics cards. Went down the path of writing a gaming system. I knew about the X86 systems, but didn't really want to get one and port Draco to it. When the Commodore Amiga came out (1984?), I got one of those and ported Draco to it. All my work was on that (and other Amiga's) for years. That system was so much better than the X86 DOS/Windows boxes of the time.

So, I went on a different path, otherwise I imagine I would have been an early Linux user, and likely contributer.

I've been on Linux for a few decades now, and wouldn't change that, but my choices haven't taken me along the contribution path.

[Denarius]

Linus in ACT

I remember the crowded discussion Linus had with Unix Group members at ANU in Canberra. That trip is when Tridge took Linux to the local Zoo where he met a penguin. Never thought the unices I worked on would be extinct or close to it from the OS Linus was talking about. I know AIX support is just moved to India, but IBMs newer Power chips dont support it, only DeadRat

[HandlesMessiah]

Missed opportunity

Linus is indirectly responsible for one of the great regrets of my life.

Shortly before he died, the great portrait cartoonist Al Hirschfeld did a massive sale of original art, and his portrait of Linus Torvalds was in the offering. I didn't get it because it was too much money (my recollection is USD3000, which in retrospect is a pittance for a Hirschfeld), and at the end of the sale, it was unsold! There was a huge rush for portraits of famous actors and musicians and politicians, and this relatively obscure nerd icon was overlooked. No idea who owns it now, but it could have been on my wall.

For reference: www.alhirschfeldfoundation.org/piece/linus-torvalds

About eight years ago I made up for it by buying three original pages from Larry Gonick's The Cartoon Guide to Computer Science that feature Claude Shannon, so I guess it all turned out okay.

[wolfetone]

It's mad to think how the tech world is today after one bloke in his dorm decided to muck around with an OS for a hobby. It's just incredible.

[Antron Argaiv]

..and shortly thereafter, the flamewar :-)

[jake]

Linux was early '90s ... Flamewars (as you understand them) started with Usenet and Fidonet, in the early '80s.

_{But I can make a case that flamewars are as old as politics and religion ...}

[Dante Alighieri]

RISC-y

Indeed, RISC had promise.

There was Acorn with RISCOS. and the ARM chip.

RISCOS is still very niche.

ARM is in the phone, telly, microwave, and increasingly the PC you use. including with Linux.

I think ARM, and RISC, has won to be honest.