Happy Birthday, Linus

Glyn Moody writes "Today is the birthday of Linus. Just under 19 years ago, on the first day the shops in Helsinki were open after the holidays, Linus rushed out and spent all his Christmas and birthday money on his first PC: a DX33 80386, with 4 Megs of RAM, no co-processor, and a 40 Megabyte hard disc. Today, the kernel he wrote on that system powers 90% of the fastest supercomputers, and is starting to find its way into more and more smartphones — not to mention everything in between. What would the world look like had he spent his money on something else?"

Comments

A case of the pundays

[suso]

How would the world look different? It would be a whole GNU world.

BTW, Linus is 40 today, there seems to be no mention of that anywhere.

Re:A case of the pundays

[dkleinsc]

Really? I wonder where you HURD that.

I hope Tove planned a big party for him, if it's the 40th.

Re:A case of the pundays

[BESTouff]

As the saying goes, Linux was an idea whose time had come. If it weren't for Linus, someone else would have invented another similar OS, or one of the existing (Hurd, BSD, Minix) would have been used. All the people, the culture, the computers and the communication medium were there, ready to give birth to such a thing.

Still, Linus has been a great leader, and a tasteful (for some) architect. Thanks a lt for that !

Re:A case of the pundays

[knewter]

I would counter that, while it inevitably would have happened at some point, it's not a given that the resulting OS would have been GPLed, and subsequently things could've turned out very differently. Happy Birthday, Linus!

Re:A case of the pundays

[Anonymous Coward]

It's GNU/Linus, you insensitive clod!

Re:A case of the pundays

[hitmark]

i guess, but linux happened at the exact right time. Hurd was a mess (i think it was at its second or third rewrite at the time, trying for the last fad in kernel design), bsd was in court, minix was anything but open/free/whatever.

so in the end, linux was a case of scratching a itch, in combo with the choice between sitting in a heated room to write code or walk across a cold campus to access the university terminals. One can say that humans are at their most creative when they want to be lazy.

Re:A case of the pundays

[larry bagina]

Minix had license/redistribution issues. It was used, but modifications could only be distributed as patches.

Re:A case of the pundays

[vegiVamp]

Unfortunately, that's the kind of saying that we can't really verify as true or false until we've invented timeline sandboxes.

Re:A case of the pundays

[SanityInAnarchy]

That's quite true -- in fact, it might have been better if it'd been under a BSD license. Linus chose the GPL almost arbitrarily, and hasn't been much for licenses...

I doubt very much that it would've been anything proprietary.

Re:A case of the pundays

[Urkki]

That's quite true -- in fact, it might have been better if it'd been under a BSD license.

I doubt that very much. I believe GPL is the reason why Linux dominates over BSD variants. It's just that more people are willing to contribute, when they feel that fruits of their labor can't be just "taken" as freely as BSD license allows.

Re:A case of the pundays

[darthdavid]

Posting to remove accidental moderation...

Re:A case of the pundays

[MrKaos]

How would the world look different? It would be a whole GNU world.

I think you mean a Brave New world where the world is a utopia of nearly clones and perfect interoperability but everyone is mentally programmed and quite insane. I'm glad it's a whole GNU world now. I suspect without Linux, BSD et. al there would be no Apple and Microsoft's monopoly would be complete.

In fact I hear that Microsoft has built a Time gateway to send a terminator back to kill Linus (Linus Torvalds? - kome wid me iv u vant to leev) except they had to engineer it using Linux which will probably cause some sort of paradox, the first one ran Windows ME and shot itself, the second ran XP and was quit stable but couldn't be convinced to 'step indo no stinkin tyme fortex' and the third ran vista and just kept banging it's head on the wall.

Wow - that post didn't end where I thought it would!!! Happy Birthday Linus.

Re:A case of the pundays

[phaggood]

Hurd was a mess (i think it was at its second or third rewrite at the time, trying for the last fad in kernel design)

Hurd is the DNF of OS's?

Re:A case of the pundays

[SanityInAnarchy]

It's just that more people are willing to contribute, when they feel that fruits of their labor can't be just "taken" as freely as BSD license allows.

Bullshit, and there are plenty of very popular projects which would demonstrate the contrary -- sqlite, for example, has no license. That's because it's entirely public domain.

To take another random example: Ruby on Rails. The license deliberately is not GPL, or even LGPL (which might have worked), but rather MIT. This means I could technically "take" it -- remember, it's not stealing, and it's not even copyright infringement here -- and build my own proprietary product.

It also means that unless I relish maintaining my own separate fork of Rails, I'll be sending patches upstream whenever I do something cool. Even monkeypatches are much easier to send in as formal patches than to maintain.

I used to think as you do, but the choice here is between the potential audience of every commercial product versus a few GNU zealots who will actually refuse to contribute to a project because they don't like the license.

I can see people contributing to Linux instead of BSD if they prefer GPL, and if there are no other factors. But if Linux didn't exist, would you really refuse to contribute to BSD?

Yesterday, I sent a patch to a project hosted by Google. They wanted me to sign an agreement essentially giving them copyright and a patent grant (without removing those rights from me) -- and this isn't Google being evil, it's common for projects to request copyright for contributions. I wasn't exactly happy about it, but it again comes down to the same choice -- are the terms of that agreement so bad that I'm going to refuse to contribute at all, or worse, fork the entire project? Probably not, especially for the small patches I have in mind.

And by the way: If you believe in the GPL, and you pirate anything (movies, music, games...), you're a hypocrite. A term common among those who have a problem with current copyright law is, "It's not theft, it's copyright infringement," implying that it's not as bad. I've occasionally heard people say that if there was no copyright, there'd be no need for the GPL, but I don't buy that -- if you really believe that, why not use BSD or MIT?

Re:A case of the pundays

[MrNiceguy_KS]

Based on their respective ages, it would perhaps be more accurate to say that DNF is the Hurd of games. Except that DNF development was powered by development dollars - which were finite, and Hurd development is powered by Richard Stallman's ideology and stubbornness - which are not.

Re:A case of the pundays

[SanityInAnarchy]

GPL is more than just about copyright. It's about giving back.

No, giving back is about giving back.

GPL is about forcing you to give back, through copyright -- or it's about restricting people who refuse to give back from using it. It was created to counter the perceived (and very likely real) threat of people building a giant proprietary product out of what was once free, and everyone inevitably upgrading to the shiny new proprietary version, leaving everyone without the ability to change their own code.

But free software, as a concept, doesn't require the GPL. Nor does free software, as a movement, rely on the GPL at this stage.

how is having a problem with one part of a law mean you have a problem with the whole law.

Copyright is a way of expressing what you want done with copies of your intellectual property.

If you have a problem with copyright, or with the idea of intellectual property, I can certainly see a case for that -- and I'd love free-as-in-beer and DRM-free movies. I can certainly see piracy inevitably crushing those who cling to draconian DRM.

But to then turn around and suggest the GPL?

Think about your reaction when you see a gpl-violations story, versus an MPAA story. When it's the RIAA or MPAA, everyone (myself included) is quick to call them the MAFIAA and to defend piracy as "copyright infringement, not theft", and even suggest that there are no moral issues with it.

But when it's a GPL violation, suddenly copyright matters and everyone is morally outraged.

They both rely on the exact same part of the law -- they both rely on the assumption that just because you wrote something, you should be able to control what people do with it.

Actually, I'll play devil's advocate for a moment -- if you consider it as a social issue, rather than a legal one, the GPL is about forcing people to share. If you think sharing is good, then you probably see the GPL as a way to encourage more sharing than would happen otherwise -- and there would probably be even more sharing with no copyright, so you'd gladly give up the GPL if it meant copyright law is gone forever. This is probably why such licenses are referred to as "copyleft".

Re:A case of the pundays

[gwern]

> I've occasionally heard people say that if there was no copyright, there'd be no need for the GPL, but I don't buy that -- if you really believe that, why not use BSD or MIT? erm... because I believe that copyright exists? '~C -> ~G' does not imply '~G'.

Re:A case of the pundays

[Kjella]

It's just that more people are willing to contribute, when they feel that fruits of their labor can't be just "taken" as freely as BSD license allows.

Bullshit, and there are plenty of very popular projects which would demonstrate the contrary

Naming two is certainly not evidence. Pick any metric you want, sourceforge license types, Linux vs *BSD commits, gcc vs whatever BSD has, anything that reasonably pits them towards each other.

I've occasionally heard people say that if there was no copyright, there'd be no need for the GPL, but I don't buy that -- if you really believe that, why not use BSD or MIT?

Copyright:
We can't take their binaries, they can't take our source.
No copyright:
We can take their binaries, they can take our source.
You:
Please bend over and give them the source anyway.

Re:A case of the pundays

[SanityInAnarchy]

You left out one of the most important projects for Linux: Apache.

I wasn't going for a comprehensive list, mainly for things I know about intimately. For example, I know that SQLite is used in tons of places it never would've seen the light of day if it was proprietary, and I'm actually grateful for that -- it means I can quite often interact with random little databases with the sqlite commandline tool, for example.

Rails is another example, for about the same reasons -- tons of webapps, probably far more proprietary than open apps, but tons of plugins and patches sent upstream, because maintaining a fork is expensive. And it is helpful in that it spreads the ideas of REST, and even if you don't know a given page is running RAILS, if it's reasonably RESTful, it's much easier to write a client for.

I really don't know any of the background behind why Apache went with the license it did.

The only convincing reason why Linux is more often used than the BSDs is that it could be used earlier... I really cannot find any other convincing reason.

Well, a derivative of that is that driver support (at the very least) is better on Linux than on other systems, but that's a product of it being popular, and it's one of several benefits that kind of feeds itself. This is the reason why, for example, I'm probably going to continue to run Linux on my fileserver and just wait for btrfs, instead of using OpenSolaris and zfs.

MySQL thrived because sometimes the GPL is unbearable and people pay money to avoid it.

I'd actually be curious to know the numbers behind that -- but I think it's again a case of popularity feeding itself. PostgreSQL has a BSD-style license, and is better in many respects -- I suspect the biggest reason it doesn't catch on is that everyone's already using MySQL, so all the apps are written with MySQL in mind, so when people actually get to the level where they want to improve a database (maybe make it clustered, for instance), MySQL is what they think of.

The main reason it seems to be bad that Oracle owns MySQL is that we haven't really seen a coherent fork come out of it that isn't owned by Oracle, and Oracle has a vested interest in not improving MySQL to the point where it could compete with Oracle's own database.

In fact I think the only reason why not every binary package for Linux must be GPLed is because FreeBSD has Linux emulation,

Not at all, and it's kind of ridiculous to suggest that.

Put another way, it's possible to have a GPL'd programming language, but the programs aren't GPL'd. Take it a step back to something familiar -- if a word processor was GPL'd, and used its own format (not ODF), would that imply every document created was GPL'd?

The GPL is actually pretty explicit about this -- the LGPL only applies to stuff that's a derivative work, and actually part of the LGPL'd library. If you link against it, but keep the original library separate, you don't have to GPL (or LGPL) the result. How separate do you have to keep them? Well, you have to basically give you users all the tools they'd need to recompile the LGPL portion -- so, maybe a .a file of all your code.

And the GPL? It applies to stuff that your program is directly linked against.

So, in what way does the Linux kernel link against any software running in userspace?

Without that every program that calls Linux kernel routines is derivative work of the Linux kernel.

Well, technically, when I make an HTTP request to a machine supporting SOAP, I might be executing a routine on that machine. If all code on that machine is GPL'd, does the software making my HTTP request have to be?

For that matter, what about GPL'd software on Windows?

No, that'd be pretty obviously absurd.

I still don't get why NVidia can distribute drivers with binary blob

Re:A case of the pundays

[SanityInAnarchy]

I suppose the point is -- if there would be no need for GPL without copyright, why wouldn't you use a license which is closer to "no copyright"?

Re:A case of the pundays

[SanityInAnarchy]

Pick any metric you want, sourceforge license types, Linux vs *BSD commits, gcc vs whatever BSD has, anything that reasonably pits them towards each other.

Firefox vs Konqueror. Even if you count Webkit, Firefox (and Gecko) still win in popularity on Windows.

jQuery vs YUI vs ExtJS. Remember ExtJS? It pretty much became irrelevant the second its author decided to switch to GPL -- the instant that happened was the instant many proprietary projects switched to other technologies. And the author was kind of a dick about it -- insisting that it applied not just to the source on the client, but to the server-side code as well.

Apache vs... wait... is there even a slightly notable GPL'd webserver?

Rails vs... um... hmm.

SQLite vs...

Seeing a pattern?

Copyright:
We can't take their binaries, they can't take our source.

Well, or they can take your source, but under conditions you impose.

No copyright:
We can take their binaries, they can take our source.

Then why not create a license that actually matches this ideal? As it is, they can take your source only if they release their source at all. They can't take your source and then give you the binaries.

You:
Please bend over and give them the source anyway.

"Bend over" and also gain developers.

Again, consider: Are you going to have more or fewer developers under GPL vs MIT? Now, there are tons of other factors, so I doubt Sourceforge is going to show much -- but it's still an interesting question to ask, especially about your software, right now.

Also consider: What do you lose if they can take your source? For all the ranting about TiVo, I'm starting to think that the invention of the DVR was a good thing (and an idea whose time had come), and it isn't as if the Linux community lost anything. If the things they needed were GPL3'd, that would probably have slowed them down significantly -- it probably would not have resulted in them allowing access to the device.

Now, all that said, I do agree with the ideals of GPLv3. I certainly would like to see more open systems. I'm just not sure this is the right way to go about it, and I'm especially skeptical that it's helpful to the individual project.

Re:A case of the pundays

[SanityInAnarchy]

Put another way, it's possible to have a GPL'd programming language, but the programs aren't GPL'd.

This really isn't that easy to answer. It comes down to what the compiler does. I guess you can write a compiler that works that way. But I'd say most are not.

Like which?

It's possible to compile proprietary programs with gcc. It's also possible to compile GPL'd programs with icc or Microsoft Visual Studio.

Another point to consider is whether there is a compiler that is not GPL'd. If I can compile that language without a GPL compiler the code cannot be considered derivative work.

Think about what you're implying here, though. If I write a Javascript app that only runs in Konqueror, is the GPL forced on me? I can't imagine that would be the case -- especially when it would be resolved simply by fixing my porting bug.

The GPL doesn't say anything about linking. That really is part of the problem. It is all about derivative work.

Fair enough. I did come across an article indicating that the problem is that there isn't a tested legal definition of derivative work, with regards to software. However...

in general I'd say if your program requires another program it is derivative work.

Two problems with that:

First, "derivative work" here is likely a legal term. Legal terms tend to be fairly well defined, but not necessarily what you'd think they are -- consider the scientific definition of "theory".

And second, if I write a GPL'd web service, and you connect to that service, I really don't see why my program should be considered a derivative work. Think about that -- Firefox can pull information directly from various Google sources. Is Firefox a derivative work of whatever Google code is running behind those services?

For me it is absolutly unclear where to draw the line.

It seems that except for a few pathological cases, it's pretty obvious -- if your program links against something GPL'd, it's covered. If it links against something LGPL'd, you're fine so long as you can produce whatever's needed to take 100% of the LGPL'd code, edit it as you wish, recompile, and link against your app.

Why do people think the GPL restricts linking only, but no other means of interaction, including OS interupts?

I'm guessing that's a longstanding convention, but OS interrupts again seem to be quite similar to making a network call -- your program is not physically connected to the OS, but rather, it runs inside the OS.

Why is it accepted to link windows libraries (like MFC), but I cannot link to other GPL incompatible libraries?

I actually have no idea. If I had to guess, I'd say what's happening here is that your GPL'd program is a derivative work of the MFC, not the other way around -- but I'm not sure how much that helps. But I'd think that actually is a violation.

Re:A case of the pundays

[gwern]

There are many cases where unilateral movement to a desired equilibrium is worse than useless; it's stupid to suggest that *obviously* copyright is not one of them. It's like saying 'if South Korea really wants peace, why isn't it disbanding its military and kicking out the US?' Because that only helps the other guy, does nothing to bring about the desired outcome, and will screw you the heck up.

Re:A case of the pundays

[Rennt]

The motivation of contributors may have been speculation on Urkki's part - but the license is the biggest difference between Linux and BSD, so it seems quite reasonable to say that the license is an important factor in Linux's comparative success.

The fact that RoR, sqlite, etc don't use the GPL is not really relevant given that they are not competing with BSD|Linux for developer|user mindshare.

Re:A case of the pundays

[SanityInAnarchy]

but the license is the biggest difference between Linux and BSD,

Actually, the sheer inertia is the biggest difference between Linux and BSD. As a developer and a user, I target Linux because it has more users and more developers than BSD -- among other things, this means it has more flexibility, because there's more ways to set up userspace, because there's just more userspace stuff in general. Plus it's going to have better hardware support, since they can't necessarily share drivers, but there are more developers for Linux.

In the open source world, popularity really can make something better, for two reasons: First, it means more developers, because every user is a potential developer -- by that I mean, some percentage of users will be developers. Windows doesn't necessarily have more developers than Linux (it may even have less?), but Linux almost necessarily has more developers than BSD or HURD.

And second, the big problems of monoculture don't apply as much -- if Windows is 90% of what I want, my only real recourse is to send Microsoft a suggestion, or switch to another OS. If Linux is 90% of what I want, I can code the other 10% -- or, most likely, someone else already has, so there will be dozens (hundreds!) of flavors, all of which have enough in common to exchange patches and ideas. So if Linux lacks something BSD has, I can actually help change that.

And the reason Linux has the marketshare is, it was in the right place at the right time -- HURD wasn't done, Minix was under a much more restrictive license (it wasn't even free-as-in-beer), and BSD was in a questionable legal state. BSD's legal status is no longer an issue, but by the time that was resolved, it was too late.

The fact that RoR, sqlite, etc don't use the GPL is not really relevant given that they are not competing with BSD|Linux for developer|user mindshare.

That is true, but if the GPL was such a huge factor, you'd expect some upstart framework/database to take over their marketshare simply by using GPL.

Re:A case of the pundays

[JackieBrown]

Is that why the BSD licensed kernel is doing so much better with proprietary drivers?

Oh wait, it's not.

Re:A case of the pundays

[jsepeta]

he should have bought a Mac :-P

Re:A case of the pundays

[snadrus]

Don't parents tell playing children to share their toys? It prevents fights, prevents the stronger winning, and allows more experiences for all.

If copyright was abolished, where would this all go? Trade secrets. Proprietary software borrowing from open software. Though source code leaks would be a lot more fun. The idea of cooperative design of works may become even more common than the Wikipedia-only common-knowledge it has today.

Re:A case of the pundays

[node 3]

There's always at least one in the bunch that fails to understand the GPL.

GPL is about forcing you to give back, through copyright

No, it's not. It's about not allowing you to redistribute a copylefted program (or derivative) without also providing access to the source.

or it's about restricting people who refuse to give back from using it.

Again, not true. In fact, I'd wager that well over 99% of people who have used GPL'd software have never, and will never, "give back."

It was created to counter the perceived (and very likely real) threat of people building a giant proprietary product out of what was once free, and everyone inevitably upgrading to the shiny new proprietary version, leaving everyone without the ability to change their own code.

Correct, but incomplete. That's sort of the straw the broke the camel's back, in a sense, but the GPL was designed to protect the four freedoms.

But free software, as a concept, doesn't require the GPL. Nor does free software, as a movement, rely on the GPL at this stage.

I think you mean open source. Free software is an established term that very much relies on the GPL, and is in fact, founded on the GPL.

This [anti-copyright sentiment] is probably why such licenses are referred to as "copyleft".

Actually, it's more to do without pointing out that while copyright is generally about restrictions, "copyleft" is about freedoms.

Whenever anyone attributes "communist" aspects to the GPL (i.e., your repeated assertion that the GPL forces one to share/contribute/give back), it speaks more to the mindset of the poster than it does any insight into the GPL. The GPL is about one thing and one thing *only*, and that is software freedom.

Re:A case of the pundays

[node 3]

I suppose the point is -- if there would be no need for GPL without copyright, why wouldn't you use a license which is closer to "no copyright"?

Because copyright *does* exist. And stating that the BSD license is somehow closer to "no copyright" is absurd, as both the GPL and the BSD License (and every other license) exists within the exact same copyright framework.

Re:A case of the pundays

[Mana Mana]

``"It's not theft, it's copyright infringement," implying that it's not as bad. ''

Your right to protest is your right to be shot in Tibet. Semantics matter dipshit.

Re:A case of the pundays

[Urkki]

It's just that more people are willing to contribute, when they feel that fruits of their labor can't be just "taken" as freely as BSD license allows.

Bullshit, and there are plenty of very popular projects which would demonstrate the contrary -- sqlite, for example, has no license. That's because it's entirely public domain.

You failed to read my (now emphasized) word "more" above.

Just think of the BSD network stack. Do you really think it's bullshit, that majority of open source crowd would be really really unhappy if their code was used like that?

Well, trust me on this: majority would be unhappy (different people for different reasons and to different degrees, but unhappy none the less). And why would they release their code under license, if it may make them unhappy, especially if their code is good enough to be used by others?

Then tehre are people who don't feel unhappy about it, or maybe feel their personal direct unhappiness is less than their happiness of indirectly benefitting mankind (even if it helps to put more cash into Bill Gates' bank account, at least some if may go to some charity of Bill's choosing...).

Re:A case of the pundays

[Urkki]

GPL is more than just about copyright. It's about giving back.

No, giving back is about giving back.

GPL is about forcing you to give back, through copyright -- or it's about restricting people who refuse to give back from using it.

Indeed, but saying "forcing to give back" is trying to intentionally put a bad spin on it. It's like saying you're forced to pay for the stuff you take from a shop. Yes, technically true, but being intentionally misleading, something a communist or a hippie might say.

Better way would be to say GPL is about making the use of open source code a fair transaction. What confuses people is claiming one or the other is more free without defining what they mean with "freedom". BSD gives the current user the freedom to forbid future users from modifying (etc) the code, while GPL forbids the current user from take freedom of modification (etc) away from future users.

And of course BSD license can be just as much about making use of open source code a fair transaction, because "fair" is subjective.

But majority of non-professional OSS contributors do feel that BSD license doesn't give them the fair transaction they want, so they don't release code they care about under BSD.

Re:A case of the pundays

[Urkki]

Isn't it ultimately matter of #included code. If you #include GPL code in your compilation, then the binary becomes GPL. Or "import" in a way that includes something from the imported thing in your code, or whatever.

If you don't include GPL data in your binary, then there's no problem. It's just that nobody maintains independent non-GPL include files for various GPL libraries, so using GPL libraries without using GPL include files is a non-issue for practical reasons.

Claiming that using an interface provided by GPL stuff forces GPL on your stuff will never hold on court. Only if your binary (or whatever package) contains bits that are derived from GPL stuff, only then it becomes GPL. If all bits of your package are of non-GPL origin, no problem.

And to clarify about output of GPL programs: FSF interpretation seems to be that output of GPL program is GPL only if it's direct copy GPL data in the GPL program. And single words (or comparable pieces of data) are too short to be under copyright, so novel combinations of words isn't under GPL either. But novel combinations of long enough pieces to be under copyright automatically make the output to be under GPL too.

And the other way too, if you pass GPL stuff through a non-GPL program that does something to it, output is GPL, because output is derived from the GPL input.

Re:A case of the pundays

[Urkki]

There's always at least one in the bunch that fails to understand the GPL.

GPL is about forcing you to give back, through copyright

No, it's not. It's about not allowing you to redistribute a copylefted program (or derivative) without also providing access to the source.

or it's about restricting people who refuse to give back from using it.

Again, not true. In fact, I'd wager that well over 99% of people who have used GPL'd software have never, and will never, "give back."

Releasing even unmodified source, and allowing users to keep modifying it is a form of "giving back". And GPL is exactly about forcing this kind of giving back. I mean, that's the whole point of the license, forcing you to grant equal rights to the users of the GPL code, even if you'd rather just release the binary containg some GPL stuff. Forcing is just a bad word for it, because it implies there's something unfair about it. Being forced to pay in a store is not unfair, so generally people don't say they're forced to pay, even though they technically are.

But free software, as a concept, doesn't require the GPL. Nor does free software, as a movement, rely on the GPL at this stage.

I think you mean open source. Free software is an established term that very much relies on the GPL, and is in fact, founded on the GPL.

FSF may try to make "Free software" mean something specific, but that's futile, because "free software" sounds exactly the same, looks almost same on paper, but (according to FSF) means something else. And telling people "please don't say 'free software' because we've reserved 'Free Software' is just as stupid as it sounds like when you say it aloud.

Actually, it's more to do without pointing out that while copyright is generally about restrictions, "copyleft" is about freedoms.

Whenever anyone attributes "communist" aspects to the GPL (i.e., your repeated assertion that the GPL forces one to share/contribute/give back), it speaks more to the mindset of the poster than it does any insight into the GPL. The GPL is about one thing and one thing *only*, and that is software freedom.

But that's ambiguous to the point of being meaningless. BSD is also about software freedom. Releasing to public domain can be about software freedom. Releasing proprietary closed source software development tools for free use can be about software freedom too.

My interpretation is, that GPL is about striving for a world where any(?) digitally copyable content is freely available for use, modification and re-distribution. It's a noble goal, and one I think would benefit mankind enormously, because of the nature of digitally copyable content. Even though it would restrict software business opportunities, it would create those opportunities elsewhere. People would have to get paid for the *act* of adding value, instead of the actual value they created. It's the same goal as copyrights and patents have, and GPL wouldn't even be needed if it weren't for the speed of current technological and cultural change, and the fundamentally different nature of digital content as compared to anything that existed when patents and copyrights were invented.

Re:A case of the pundays

[Urkki]

But what you say would make the GPL quite weak. I could just write my own header file I include. The function names and argument types have to match. But that is what we call the API and an API itself can't be protected with copyright. Then I just have to link dynamically and then there are no GPL bits in the binary.

While I think that would be a rational and usable description of "derivative work" I don't think this is what the people using the GPL for their programs have in mind.

It doesn't matter what people have in mind, what matters is what is written in law and what is written in the license.

What you describe is not an issue., because it's too much work, too hard to maintain and distribute. And it'd be useless for embedded devices, since you'd have to distribute the GPL library in the device anyway, in a way that it's automatically combined with your code.

It would work for a PC software, where you distribute the binary blow that was compiled with your own #include files, so that your distribution does not include any GPL code. There's nothing in the binary that is under GPL, so GPL simply does not apply to it. How could it?

Re:A case of the pundays

[ramana8]

Happy Birthday Linus. Thanks for a nice operating system. I use it everyday. Alas, I still make a living on that other operating system. :-)

Re:A case of the pundays

[SanityInAnarchy]

You failed to read my (now emphasized) word "more" above.

I read it, I just disagree.

Just think of the BSD network stack. Do you really think it's bullshit, that majority of open source crowd would be really really unhappy if their code was used like that?

Yes.

First, think of one of the primary motivations of open source, or of software development, or of any creative endeavor to begin with -- you want your creation to be appreciated. You want your code to be used.

But also, think of the BSD stack itself -- doesn't it make sense that this, of all things, should be free to reimplement? Certainly, any new network protocol or document specification should have a commercially-reusable implementation. Or, to put it another way: Would you rather we have the BSD stack re-used everywhere, or would you rather people keep inventing weird stuff like IPX?

And why would they release their code under license, if it may make them unhappy, especially if their code is good enough to be used by others?

First, I think there's a subset who actually don't realize it's an option to use something else, or haven't really thought about it thoroughly.

I also think there are a fair number who actually would be upset, as you've said. Certainly, there are projects I would hate to see lifted wholesale, rebranded, and released without so much as crediting me. And the fact that people do get outraged at GPL violations would tend to support your point.

But a large part of why people get outraged at GPL violations is simply the fact that the open source community was very clear about how they wanted their code to be used, and that's being exploited. I'm generally really unhappy when people violate the GPL, even if I don't always think the GPL should've been used in the first place.

There's also the subset who release code under the GPL because it's required by the project they're committing to, or because they're using some other GPL'd library -- maybe they wanted Qt, for example, before it went LGPL.

Then tehre are people who don't feel unhappy about it, or maybe feel their personal direct unhappiness is less than their happiness of indirectly benefitting mankind (even if it helps to put more cash into Bill Gates' bank account, at least some if may go to some charity of Bill's choosing...).

That's not it at all, and I have no particular love of Gates, nor do charities help that.

But I'd rather Windows support IP properly, even if they have to use my code, than see them decide to invent their own. If IE ever does support HTML5 local storage, I'd rather see it embed SQLite than see them try to build it on Access or something.

And that's assuming it's Microsoft who uses it. It could always be a startup who wants to challenge Microsoft. I like that some of my code might make it easy for some small startup to go from zero to competitive -- that it lowers the barrier of entry, that it makes it a more level playing field.

There's also the extent to which I care much more about actually building software than what license it's under. I have no problem releasing GPL code. The biggest reason I appreciate licenses like BSD and Apache, though, is that I can pretty much glance at the LICENSE file, or even read through all of it (it's tiny!), and then just get to work. I don't have to think about whether it's GPL, or whether the code I want to use is GPL-compatible, or how far the LGPL extends, or whether I really want to trust the FSF by using "any future version" (or if, conversely, I want to force contributors to cede copyright to me so I, or the project as a whole, can choose to switch licenses).

And it's far easier to go from an MIT license to the GPL than the other way around -- if I find a large number of developers really do want to use the GPL, I don't need anyone's permission to incorporate the old MIT code.

Re:A case of the pundays

[SanityInAnarchy]

No, but that's why, for example, it's the BSD-licensed kernels that have ZFS, which actually cannot be incorporated into the Linux kernel, for legal reasons. They're both open source, but the GPL puts up this wall of legal bullshit, so they can't actually exchange code.

Re:A case of the pundays

[SanityInAnarchy]

I think you mean open source.

No, I don't.

Free software is an established term that very much relies on the GPL, and is in fact, founded on the GPL.

No it's not.

Free software, as you've described, is software which provides the Four Freedoms. BSD-style licenses indeed provide these freedoms, as does the public domain.

What it doesn't do is force any derivative work to also provide the same freedoms. This means that someone could create a project which is derived from a BSD-licensed project, but which is proprietary and does not offer those freedoms -- but that fork would also no longer be BSD-licensed. The original code, under the original license, is still Free as in freedom.

But I don't think I have to say much more when the FSF itself disagrees with you.

Whenever anyone attributes "communist" aspects to the GPL (i.e., your repeated assertion that the GPL forces one to share/contribute/give back), it speaks more to the mindset of the poster than it does any insight into the GPL.

I never claimed it was communist, nor do I necessarily believe communism is bad. Your kneejerk reaction to any perceived kneejerk reaction is especially ironic.

Re:A case of the pundays

[SanityInAnarchy]

My interpretation is, that GPL is about striving for a world where any(?) digitally copyable content is freely available for use, modification and re-distribution.

That is probably true. Unfortunately, in the short term, it has some nasty side effects of occasionally taking a step back from that goal -- for example, Linux can't use ZFS, because the licenses are incompatible, and specifically because of the use of the GPL on the Linux side.

Re:A case of the pundays

[SanityInAnarchy]

Indeed, but saying "forcing to give back" is trying to intentionally put a bad spin on it.

My intention wasn't to spin it...

It's like saying you're forced to pay for the stuff you take from a shop.

Right. I've generally argued the other side of this -- it's a fair transaction, to say that if you take my code, you should give something back, or you should obey my wishes. I've gone so far as to argue in favor of GPLv3, or even something stricter.

One point I've made, though, is that we have this idea of "taken" -- we see plenty of people outraged when it's the fruit of their labors which are used against their will. Yet on the other hand, we see people who dislike copyright, who pirate and feel it should be legal, and we see the GPL's goal as being a world without copyright, where any digital media is free to use and remix.

But these are contradictory.

Re:A case of the pundays

[SanityInAnarchy]

stating that the BSD license is somehow closer to "no copyright" is absurd, as both the GPL and the BSD License (and every other license) exists within the exact same copyright framework.

That is true, but it is quite clear that the BSD license imposes fewer restrictions on the user. No copyright imposes no restrictions on the user. By that metric, I claim fewer restrictions is closer to no restrictions.

Re:A case of the pundays

[SanityInAnarchy]

Fact: OS X is not dead. OS X is BSD.

More relevant fact: This is about the BSD license, and BSD-style licenses like MIT. It has little to nothing to do with the BSD software package.

Re:A case of the pundays

[suffe]

I suppose that the point is -- if there was no copyright, there would be no need for the GPL because you wouldn't have a leg to stand on.

From my, admittedly limited, knowledge of the GPL it stems from copyright law and not contract law; i.e. given the limits imposed on you due to a copyright claim on a certain software you are granted ADDITONAL rights. Thus, no copyright, no need for the GPL.

On a slightly less legal and more philosophical plane, the GPL is there to prevent something once free from turning non-free and thus limit people from using it as they want. Without copyright you could just 'liberate' it again. Disassemble and rewrite if need be. The focus on 'contributing back' takes a back seat to this.

Re:A case of the pundays

[SanityInAnarchy]

the GPL is there to prevent something once free from turning non-free and thus limit people from using it as they want. Without copyright you could just 'liberate' it again. Disassemble and rewrite if need be.

The GPL goes further than that, though. If that was all, it would be legal to include GPL in freeware, so long as you didn't restrict people from disassembling it.

As it is, you are required to provide source. Obviously, someone thought "liberating" it through disassembly wasn't enough.

Re:A case of the pundays

[lawpoop]

I would counter that, while it inevitably would have happened at some point, it's not a given that the resulting OS would have been GPLed, and subsequently things could've turned out very differently.

I disagree. The very idea whose time had come was a GPL'ed operating system, leveraging *ahem* the open nature of the internet and collaborative tools.

There were some also-rans during that era, such as BeOS and all the other proprietary Unix clones. But the reason that they were also-rans is because they were closed -- they could not be freely spread, and therefore could not compete with Windows, which had a giant organization devoting much of its resources towards the propagation of Windows.

So long as there wasn't a GPL'ed OS, there would have been a serious hunger for it. *Something* would have arisen.

Re:A case of the pundays

[diego.viola]

You are wrong, GNU would be nowhere if it weren't for Linux..

Re:A case of the pundays

[True Grit]

there are plenty of very popular projects which would demonstrate the contrary

None of your contrary examples are operating systems.

but the choice here is between the potential audience of every commercial product versus a few GNU zealots

You forgot the other group that has become very significant to the relative success of Linux: commercial companies that use Linux (and contribute code to it) because they know their competitors can't take anything they contribute, put it in a closed-source alternative, and use it against them. Meanwhile, they all get the advantage from the code contributions of others. Its a more level playing field for many commercial entities, and they don't have the hassle of maintaining their own private fork.

It all depends on the nature of the market. For normal apps, the license issue is not that significant (except to the zealots - on whichever side), but, as an example, for commercial companies who need an embedded OS for something (where the OS itself is not the center-piece of their product - thus the fact that their competitors can use it too is irrelevant), the license really does make a difference, and many of them are deliberately choosing the GPLed Linux (for completely agnostic, practical reasons).

The point I'm trying to make here is that the license issue is actually important (in both directions) to a *lot* of people/companies, not just the 'zealots' (of either side).

But if Linux didn't exist, would you really refuse to contribute to BSD?

No, but then again, if Linus hadn't of done it, it would only have been a matter of time before someone else released a GPLed OS kernel, and once that happened, I believe the same results would have happened as well: it would have become popular despite the pre-existence of the BSDs, because for some people, and for some companies in some markets, the GPL makes better sense for them.

You sound like you desperately want to believe the only reason the GPL and Linux is so widespread is because of zealotry, but I don't see that, there just aren't that many GPL zealots out there (they're obviously badly outnumbered here on /.). RMS's army of smelly, bearded holy warriors isn't *that* big. :)

There really is more going on here than just zealotry...

Re:A case of the pundays

[True Grit]

One point I've made, though, is that we have this idea of "taken" ... on the other hand, we see people who dislike copyright, who pirate and feel it should be legal

Thing is, you're the only one who keeps bringing this point up, it was never part of the original issue, and the post you initially responded to said nothing about this.

Now here's the problem you have: there are plenty of people who agree with your point above... but also use the GPL. To them, you're starting off on a GPL rant, and ending up on piracy/MPAA/whatever, is merely going off-topic, not making a point.

I'd suggest that you actually verify your implicit assumption that everyone who advocates the use of the GPL also advocates piracy, and once you discover this assumption is false, save the second half of your rant just for those to whom it actually applies.

Re:A case of the pundays

[SanityInAnarchy]

I'd suggest that you actually verify your implicit assumption that everyone who advocates the use of the GPL also advocates piracy,

I never explicitly said that, nor intended to imply it.

save the second half of your rant just for those to whom it actually applies.

Didn't I already do that? Here's what I said:

If you believe in the GPL, and you pirate anything...

Obviously, if you don't, that doesn't apply to you.

Re:A case of the pundays

[SanityInAnarchy]

None of your contrary examples are operating systems.

Is your point, then, that the GPL is crucial to the success of an operating system? What is special about operating systems that makes this the case?

You forgot the other group that has become very significant to the relative success of Linux: commercial companies that use Linux (and contribute code to it) because they know their competitors can't take anything they contribute, put it in a closed-source alternative, and use it against them.

Of course they can. Google maintains a proprietary fork of Linux, without violating the GPL, because they don't distribute, but they do try to keep pace with actual kernel development.

The nature of the open source development model is what encourages companies to contribute back, because there is a very real cost in doing what you describe, especially for an established open project. I think I've discussed this already -- any changes you keep private, you have to maintain, and this maintenance overhead is doubled if you want to actually track the public development. You're basically merging all the time.

This is the same cost as the cost of forking in the open source world, only more so, because forks can at least still see what each other are doing, maybe exchange code both ways, maybe even merge at some point.

Now, there is a group I'm neglecting -- the dual-licensing group, companies like MySQL and TrollTech, who have a business model that quite literally relies on people being willing to pay for a license that lets them do things the GPL won't. But even here, TrollTech now distributes Qt under the LGPL, meaning the only people who would need to commercially license it are people who make proprietary changes to Qt itself. I would guess they get more from support than from that.

There's probably some overlap between this group and the group you described -- for example, Quake 3 is released with the assumption that if your competitors are going to use it, they're either going to be open source (so you can see what they did), or they're going to pay you for a different license. I suspect this is where you'd have your strongest argument -- though I also suspect that licenses like the Mozilla Public License will be common.

as an example, for commercial companies who need an embedded OS for something (where the OS itself is not the center-piece of their product - thus the fact that their competitors can use it too is irrelevant), the license really does make a difference, and many of them are deliberately choosing the GPLed Linux (for completely agnostic, practical reasons).

Right -- agnostic, practical -- if you mean what I think you mean, they aren't choosing it for the license, they're choosing it because it's stable, mature, and does what they want.

Note that this argument would also apply to open source video codecs, standard libraries, GUI toolkits... I doubt there are many embedded systems which are complete except for a kernel.

if Linus hadn't of done it, it would only have been a matter of time before someone else released a GPLed OS kernel, and once that happened, I believe the same results would have happened as well

I don't see how you can look at the history and conclude that.

That is: Yes, there would have been a GPL'd OS kernel. By the time that happened, BSD would probably already have filled the niche pretty thoroughly. As you've said:

for some people, and for some companies in some markets, the GPL makes better sense for them.

And for other people, and other markets, the GPL makes less sense, or is a non-starter, so I don't really see what that proves.

You sound like you desperately want to believe the only reason the GPL and Linux is so widespread is because of zealotry,

Then let me clarify:

The reason Linux is so widespread is not

Re:A case of the pundays

[True Grit]

None of your contrary examples are operating systems.

Is your point, then, that the GPL is crucial to the success of an operating system?

Only that an OS and an app have different characteristics when it comes to the license issue. First, OSes are far more complex. Second, for gadget makers who need an embedded OS, the OS is just the foundation of a software stack, and its usually 'hidden', so gadget makers will have less of an issue with using an OS that their competitors can also use (and see). For those 2 reasons the enforced collaboration/sharing that the GPL requires makes more sense for some companies/people.

Of course they can. Google maintains a proprietary fork of Linux

My previous/current comments were/are referring to code that is distributed, thus my examples involving companies using an embedded OS on consumer products.

Besides, private forks are irrelevant to this discussion since they don't need a license anyway.

The nature of the open source development model is what encourages companies to contribute back,

Encourages, but doesn't require, unless the GPL is used. Some companies will not consider it acceptable to risk contributing if their competitors are not also required to accept the exact same risks (and rewards). Thus a level playing field.

as an example, for commercial companies who need an embedded OS for something (where the OS itself is not the center-piece of their product - thus the fact that their competitors can use it too is irrelevant), the license really does make a difference, and many of them are deliberately choosing the GPLed Linux (for completely agnostic, practical reasons).

Right -- agnostic, practical -- if you mean what I think you mean, they aren't choosing it for the license, they're choosing it because it's stable, mature, and does what they want.

No, they're choosing a GPLed OS rather then a BSDed OS precisely because of the license, but not for the usual 'religious' reasons. Again, as above, its a level playing field.

I don't see how you can look at the history and conclude that.

We're talking about a hypothetical here, the history doesn't help us.

And for other people, and other markets, the GPL makes less sense,

I never claimed the GPL works for everyone, only that it works better for some.

Then let me clarify:
The reason Linux is so widespread is not because of zealotry.

Thank you. The whole 'GNU zealots' thing is getting real old, especially since real 'GNU zealots' are as rare as hen's teeth around here.

It's because by the time anyone else (who wasn't BSD) had figured out that this might be a good idea, Linux had already snowballed into the default choice.

This is where we (still) disagree. I don't believe the BSDed OSes would 'win' just because they arrived first, and by the same reasoning I believe Linux is not as popular as it is just because it arrived first.

So as soon as our hypothetical GPLed OS did become available, those companies/people for whom the GPL works better would migrate to it because it *does* work better for them, no matter how long the prior BSDed OSes had been around.

If one license is universally better than the other one, in all circumstances, then we'll inevitably see a movement to that one 'superior' license. The fact that that doesn't appear to be happening tells me that the licensing does matter, and that the different licenses are useful to different people (or the same folks but in different circumstances).

While I can see the attractiveness of the GPL for a (very) young project, it seems like it's more about wanting to put _some_ license on it so you can get back to the part that counts -- actual code.

Re:A case of the pundays

[SanityInAnarchy]

First, OSes are far more complex.

...than what? Remember, we're talking about an OS kernel. Especially if you're talking about embedded systems, what they're going to do first is strip out everything they don't need -- which is a lot.

Second, for gadget makers who need an embedded OS, the OS is just the foundation of a software stack, and its usually 'hidden', so gadget makers will have less of an issue with using an OS that their competitors can also use (and see).

Well, yes and no -- it's still bitten them in the past. For instance, Linux was used in a Linksys router, which forced them to release some code (though mostly blobs) related to the Broadcom drivers. The more innovative your gadget, the more likely you'll have to write a driver, and the more likely you might be to want to keep that driver secret -- so it again points to BSD winning.

The reason Linux dominates this space, I think, is because there are that many examples of Linux working, and because Linux has that much momentum elsewhere -- we know it's reliable, and there are probably existing products we can buy to make things easier on ourselves. This isn't the case for BSD, but it does seem like a better license for the "gadget maker".

private forks are irrelevant to this discussion since they don't need a license anyway.

They are very relevant, because they are a "loophole" in the GPL, and because most people don't seem to be bothered by them.

But if you insist, consider nVidia. They seem to have found a loophole which lets them insert their own code into the kernel. At that point, it seems to me that the Mozilla license makes more sense -- if you can violate the process boundary anyway, it at least "protects" individual source files.

Some companies will not consider it acceptable to risk contributing if their competitors are not also required to accept the exact same risks (and rewards).

The only difference here is that they are given a choice between the risks (and rewards) of contributing, and the risks (and rewards) of maintaining a separate branch. All such a company has to do is look at the costs they themselves would face by maintaining a separate branch.

It's also worth mentioning, though this is sometimes a hard point to make to a corporation: Your secret sauce is not Linux. Nor is it, say, Ruby on Rails. If those were the only things making you stand apart from the crowd, you wouldn't have a product.

Take Twitter, Livejournal, etc. They've both released open source projects. Not just contributing back when they don't have to, but releasing a brand-new project (memcached, redis, etc). But when you think about it, what sets twitter apart isn't their ability to scale -- indeed, when they have issues (the Fail Whale), that doesn't really seem to hurt them.

We're talking about a hypothetical here, the history doesn't help us.

Given the history, the entire motivation for Linux was that nothing like it existed -- if PC-BSD was usable, Linus probably wouldn't have written it.

So if you're a potential developer at the time, do you start a brand-new project, or do you contribute to BSD? If you see a fully-formed BSD and a fledgeling GPL'd kernel, which do you contribute to?

The only reason I can see wanting to contribute to something else, like HURD, is that it is technologically better, in theory -- I'd really like to see a good microkernel system. But HURD still hasn't made much progress.

That's a similar situation -- if HURD had been released, Linus probably would've contributed to it. Indeed, he more or less assumed (early on) that Linux was a stopgap until HURD was ready. But as I understand it, HURD was about as stalled then as it is now.

Basically, what you're claiming, hypothetically, is that something completely different -- not Linux, not HURD, and not BSD -- would've been invented, and woul

Re:A case of the pundays

[True Grit]

For instance, Linux was used in a Linksys router, which forced them to release some code (though mostly blobs) related to the Broadcom drivers.

It was a binary driver, they were not forced to release the source code to it, only the source to the rest of the FOSS software they were using.

The more innovative your gadget, the more likely you'll have to write a driver, and the more likely you might be to want to keep that driver secret

Which you can do with Linux.

private forks are irrelevant to this discussion since they don't need a license anyway.

They are very relevant, because they are a "loophole" in the GPL

Not being applicable to internal usage was a deliberate intent of the GPL, it is only concerned with distribution to the larger public.

But if you insist, consider nVidia. They seem to have found a loophole which lets them insert their own code into the kernel.

??? That is not a 'private fork'. You seem to have changed subjects again...

The only difference here is that they are given a choice between the risks (and rewards) of contributing, and the risks (and rewards) of maintaining a separate branch.

Thats obvious. My only point is that for some companies the risk/reward of a GPLed OS makes better sense for them.

Basically, what you're claiming, hypothetically, is that something completely different -- not Linux, not HURD, and not BSD -- would've been invented,

Not completely different in the technical sense, only in its licensing.

and would've surpassed BSD, which would clearly dominate, at least for a short time.

Note that I never said 'surpass' or 'dominate'. Only that those for whom the licensing works better would switch to it eventually. Just as those for whom the BSD works better already have been, or will be in the future, switching to the BSD over time.

And that is where we differ, because I really don't see that happening.

Fair enough. That part of my post was a 'hypothetical of a hypothetical' for which I have no evidence.

for example, the reason projects like Postfix are successful is because

You keep going back to non-OS examples, when my point was all about licensing issues related to operating systems...

And you keep ignoring my point that neither the GPL nor the BSD work for everyone, so obviously there will be plenty of examples for each side...

KHTML probably never would've made it anywhere close to Gecko

Becoming Webkit was never its goal, it was just KDE's internal HTML renderer, which was then taken by *others* and turned into something more than even its original designers had intended.

So you're right, this is where we do differ -- because BSD does work better for other companies,

Sigh... when did I ever say the GPL worked for everyone? All I said was the corollary of your above point is also true. Thats why the GPL and BSD will always both be around.. because neither one works best for everyone.

What I'm arguing is that there are plenty of people who choose Linux today, even though it's not the license they'd prefer,

If there are enough of those people to form a large enough dev community for the BSDed OSes, then the BSDs will eventually catch up to Linux, and those people will switch.

My point is only that you shouldn't expect Linux to die at the point, since, for some/many people, its licensing works better for them (again, a corollary)...

I realize you're claiming it wouldn't get that far.

Not at all, you keep making the mistake of assuming I'm one of those 'GNU zealots' who hates the BSD and wants to see it die.

Re:A case of the pundays

[SanityInAnarchy]

The more innovative your gadget, the more likely you'll have to write a driver, and the more likely you might be to want to keep that driver secret

Which you can do with Linux.

Smells like a GPL loophole, then.

??? That is not a 'private fork'. You seem to have changed subjects again...

That is: "if you insist they are irrelevant", here's another example of a loophole in the GPL.

Note that I never said 'surpass' or 'dominate'. Only that those for whom the licensing works better would switch to it eventually.

No, but you did say this:

I don't believe the BSDed OSes would 'win' just because they arrived first, and by the same reasoning I believe Linux is not as popular as it is just because it arrived first.

You definitely seem to be implying that Linux succeeded more because of its license than the fact that it arrived first.

Alright, then, what do you think would've happened, if BSD had been available? I already know you predict a GPL'd kernel would emerge, but what do you think would be different?

Becoming Webkit was never its goal, it was just KDE's internal HTML renderer, which was then taken by *others* and turned into something more than even its original designers had intended.

No, it seems quite obvious, based on the fact that Konqueror exists, that its original designers had always intended for it to be a fully-functional HTML renderer. That's the part where Webkit succeeded, and KHTML (pre-Webkit) failed. There was definitely a time when replacing KHTML with Webkit in Konqueror would've been an improvement.

But I'm not sure the goal is relevant here. The point is that it was very highly successful, and it was so successful because of its technical merits, not because of its license.

Sigh... when did I ever say the GPL worked for everyone?

I wasn't accusing you of that. Please, read the rest of the paragraph. At least read the rest of the sentence:

BSD does work better for other companies, but that hasn't resulted in significant progress in BSD...

That is, in BSD, the OS. The point I'm arguing (which I'm no longer sure you disagree with), is that the success of Linux is due mostly to the size of its developer base, which is due mostly to its head start -- that the license is completely secondary to both of those factors.

My point is only that you shouldn't expect Linux to die at the point, since, for some/many people, its licensing works better for them (again, a corollary)...

I never said it would.

And yet C has been eclipsed by many other languages since then, in terms of 'popularity'. On the other hand, its still around and being used because its still useful to some people. Really, I don't understand your point here.

Where are you getting this? It seems like C is still as popular as ever. Most of the languages which have come even close to it are indeed written in C, as are OS kernels, despite better technologies existing.

On Unix, at least, all language bindings and OS interaction tend to go through C at some point.

My point here is, quite simply, worse is better. Looking at the contenders, there's C++, which has most of the same flaws as C; Java, which is basically a castrated C++; C#, which is basically Java with some useful stuff added; and PHP and VB, which are both paper-bag ugly.

The best language that's even close to C is Javascript.

Where's Lisp in all this? Where's Erlang, or Haskell? There are dozens of other languages that would be better choices. The article is specifically about Lisp vs C, and yes, I would have preferred Lisp won.

So my point is, having something done first, and having it be simple to implement and improve, means it will acquire many more developers than the Right Way. HURD

Re:A case of the pundays

[mannd]

You mean GNU/Linus, don't you?

Debian GNU/kFreeBSD

[tepples]

What would the world look like had he spent his money on something else?

Not much different, as the people who built Linux distributions would instead have ported GNU to the kernel of FreeBSD.

Re:Debian GNU/kFreeBSD

[eln]

There's also at least a small chance that many of the kernel hackers who work on Linux today would have been working on the Hurd kernel. As it happened, the release of Linux essentially killed Hurd, although it's technically still around.

Re:Debian GNU/kFreeBSD

[Pharmboy]

While it would run in similar POSIX way, there is enough difference in the BSD and GNU/Linux license, that Free software would be very different. Don't underestimate the power of the philosophy behind the software.

Re:Debian GNU/kFreeBSD

[Anonymous Coward]

If Bill Jolitz hadn't dropped off the face of the Earth for a year and perhaps even when he wasn't incommunicado if he had been more receptive to help from other people who wanted to pitch in we might be running a lot more 386BSD.

Instead he ceded the high ground (IMO) to Torvalds.

Re:Debian GNU/kFreeBSD

[Anonymous Coward]

and there would have been a Debian project without Linux? would the GNU-utils have been where they are today without the massive attraction to them through Linux? I'm not saying they wouldn't have improved without Linux, but they sure as hell wouldn't be as good as they are today.

Re:Debian GNU/kFreeBSD

[Anonymous Coward]

For at least the past few years, probably longer, Hurd has been in a kind of limbo state where the old core technology (Mach kernel) is hopelessly outdated, the new-but-stable (L4) isn't very interesting, and the really cool ones which are capability-based (L4.sec, Coyotos, Viengoos) were still in development...as far as I can tell, even Viengoos has been more or less abandoned now. Hurd is effectively dead and/or pointless until we have a good capability-based microkernel. Which would be a huge innovation for an OS that might be usable by people other than researchers.

Re:Debian GNU/kFreeBSD

[LOLLinux]

As it happened, the release of Linux essentially killed Hurd, although it's technically still around.

Hurd killed itself by the same means that DNF was killed. Linux just helped speed up the process.

Re:Debian GNU/kFreeBSD

[dkf]

There's also at least a small chance that many of the kernel hackers who work on Linux today would have been working on the Hurd kernel.

I doubt it would have happened. The Hurd hackers wanted to do fundamental OS research, and everyone else wanted a "Unix" kernel that they could just use and hack around with, and which didn't cost a lot.

I can remember that the biggest factor in our little group of hackers moving to Linux (from 386BSD) was that it had working shared libraries. OK, they sucked in many (many!) ways, but it still meant that you didn't need to have loads of copies of libc in memory or on disk at once. On the small machines of the time, that was a massive saving.

Re:Debian GNU/kFreeBSD

[Himring]

A caveman built a wheel. This other caveman had been trying to build it first, but, alas, didn't. So, now, we have a wheel. It's not just that there's no need to build yet another wheel, it's that, philosophically, wheel is built....

Re:Debian GNU/kFreeBSD

[Anonymous Coward]

...didn't need to have loads of copies of libc in memory or on disk at once. On the small machines of the time, that was a massive saving.

Oh how the times have changed... now we run entire VMs just to run single applications...

Re:Debian GNU/kFreeBSD

[FictionPimp]

What I still don't understand is why BSD didn't fill the void that linux filled. Just because of the permissive license?

Re:Debian GNU/kFreeBSD

[nabsltd]

Oh how the times have changed... now we run entire VMs just to run single applications...

And yet, there is still similar sharing going on.

Almost every VM hypervisor (and definitely every good one) uses a shared code page system so that there aren't duplicate copies in RAM. Recently, I was running 4 Fedora VMs set to 512MB of RAM each, and the total memory the host used was about 600MB.

Re:Debian GNU/kFreeBSD

[LOLLinux]

Or maybe because it was mired in legal ambiguity for years? Why would a permissive license have hurt its adoption?

Re:Debian GNU/kFreeBSD

[icebraining]

Because BSD was being disputed in court:

The lawsuit slowed development of the free-software descendants of BSD for nearly two years while their legal status was in question, and as a result systems based on the Linux kernel, which did not have such legal ambiguity, gained greater support. Although not released until 1992, development of 386BSD predated that of Linux. Linus Torvalds has said that if 386BSD had been available at the time, he probably would not have created Linux.

BSD 4.4 was only released in '94, due to the lawsuit, and Linux had already gained popularity.

Re:Debian GNU/kFreeBSD

[onefriedrice]

I know people like to think that the GPL played a significant role in the success of Linux, but in the face of so many examples of successful non-GPL free software such as apache and postfix, perhaps that assertion should be accompanied by a reasoned argument. Since no single software license is clearly a common denominator for successful free software, I wouldn't assume that the GPL is the driving force behind the success of Linux unless somebody makes a convincing argument in support of that claim. As it is now, although the assertion is often repeated, it seems to be based on mere assumption.

Re:Debian GNU/kFreeBSD

[Pharmboy]

Of course the GNU/GPL is part of the success of Linux. On one extreme, there are those who believe GPL is more than a license, it is religion itself and RMS is their prophet. If the license didn't matter, many of the people writing software only for the GNU/GPL now might have working on BSD instead. Not every programmer is license agnostic. That doesn't mean that Free software (as a whole) would necessarily be behind the current state, but there are many people who are are a part of the "Linux only" scene *solely* because they believe in the idea of the GPL. For some, "free" isn't good enough, it must be "free, with the obligation it stays free". Some of those people just so happen to be pretty good programmers.

Re:Debian GNU/kFreeBSD

[SanityInAnarchy]

If you mean VM as in Xen or VMWare, see natbsltd's post.

If you mean VM as in Java, Python, Ruby, etc, there's a few things to consider: The cache is shared as before (so there aren't duplicate copies of the VM), and you can run multiple apps in the same VM.

Re:Debian GNU/kFreeBSD

[Bert64]

Lots of companies have contributed to Linux, many of them would most likely have created closed source forks of BSD because of the permissive license (see nokia ipso, juniper etc)...

Re:Debian GNU/kFreeBSD

[Bert64]

Where many companies have contributed to Linux, their efforts have been available to everyone...
Many companies have taken BSD variants and produced closed source forks, with the results not being shared with the community. There are a large number of commercial products out there based on modified BSD...

Re:Debian GNU/kFreeBSD

[chrb]

Just because of the permissive license?

Partly - commercial BSD derivatives and the BSD networking stack ending up in Windows 95 put off some developers - who wants to see their work being co-opted by Microsoft and other corporations in closed source products? Some people no doubt, but not the majority. Another effect of the licensing was that BSD splintered into different, slightly incompatible commercial forks. The GPL protected Linux from that fate - the free distributions were shipping more-or-less exactly the same kernel code as the commercial distributions - the general perception was that with Linux free didn't mean "less", it just meant "no telephone support". In BSD land, "free" (to many) meant "not as good as what Sun and IBM are selling".

Another factor was that BSD was seen by some Europeans as being controlled by some American labs and American universities. This made it seem less approachable - and harder to get your code into. Linus, in contrast, welcomed good patches with open arms. Linus was highly enthusiastic about people developing code for his kernel. The same enthusiasm for outsiders was not visible within the BSD community.

Yet another factor was the number of Linux distributions that sprung up. Competition is good - Debian, Slackware, Red Hat, etc. were all competing to make the best Linux distribution, and there were numerous other distributions trying to push them from the top. In contrast, BSD was more centrally controlled, and whilst there was some competition between distributions, there wasn't a great amount. Plus the licensing made forks more likely - with GPL and Linux, if someone wrote a good patch, it was highly likely that patch would end up in all Linux distributions fairly rapidly. The same could not be said of BSD and its various commercial forks.

Re:Debian GNU/kFreeBSD

[Bert64]

That only accounts for general purpose OS releases...

Juniper routers use a derivative or NetBSD
Nokia firewalls run the FreeBSD derived IPSO
Many commercial appliances are based on BSD too...

Re:Debian GNU/kFreeBSD

[operagost]

That being said, I doubt that a caveman invented the wheel. The oldest wheel we've discovered is from 3,500 BC.

Re:Debian GNU/kFreeBSD

[shoor]

In the comments to a recent slashdot topic on Duke Nukem, somebody mentioned that the perfect is enemy to the good. In that context, it was suggested that the desire to make Duke Nukem 'perfect' is what killed it. I'm too much of an outsider to know for sure, but I can't help but suspect that the desire for perfection in the Hurd was a major difficulty. When something 'good enough' came along, that was just the straw that put it in limbo.

Re:Debian GNU/kFreeBSD

[chrb]

The GPL strongly discourages proprietary forks - if somebody writes great code, anybody receiving that kernel can get that code and send it back upstream. There are proprietary forks of Apache - Apache is successful *despite* these forks - if it were GPLed, then the code that Oracle and IBM roll into their own Apache-based products would be much more likely to end up being contributed back upstream. One of IBM's Linux managers once made the point that it was the GPL that forces companies like IBM, Sun, Novell and Red Hat to work together on the kernel without having to spend months negotiating technology cross-licensing and transfer agreements for every single patch. The legal basis of the GPL is solid, and these large companies understand that they must release the source to their GPL derivative works. Once that is understood, there is no reason not to cooperate - there is far more to be gained by cooperating than by not.

This diagram shows the cost of not cooperating - note the number of forks of BSD, and then note the very simple vertical history of Linux. It was the GPL that ensured the Linux of 1991 was the same as the Linux of 2009.

(That is obviously not the full story - the other major factors can be put down to Linus turning out be a surprisingly good project manager, and the emergence of internet access at home, which enabled a new generation of hackers outside of university and the large corporations to begin contributing to open source projects).

Re:Debian GNU/kFreeBSD

[LOLLinux]

Partly - commercial BSD derivatives and the BSD networking stack ending up in Windows 95 put off some developers - who wants to see their work being co-opted by Microsoft and other corporations in closed source products?

Only GNU zealots? The people behind the original BSD, and the modern day ones, were/are perfectly fine with commercial companies using their code in their closed-source products.

Re:Debian GNU/kFreeBSD

[bytesex]

What won me over for Linux, because I dabbled both in Linux and FreeBSD at the time, was the fact that, ultimately, Linux had an 'easy' GUI (startx) and.. *drumroll* Netscape! BSD only had Linux compatibility mode (after a few more years) and 'make world'. It is with Linux as it is with Microsoft: you should never underestimate the power of the half-geek. Especially in great numbers.

Re:Debian GNU/kFreeBSD

[Rob Riggs]

That's only partly true. In '94 BSD was even or ahead of Linux in terms of features. The reason Linux ended up "the winner" is because there was a stark difference between the two communities in welcoming newbies into the fold. #unix was the place to go on IRC for abuse. In stark contrast, the folks on #linux were very patient and helpful.

I had both 386BSD and Slackware downloaded to floppy. I ended up running Linux because I was welcomed by the Linux community. Not so much with the BSD crowd. A little kindness is all it took to make Linux the world's most popular Unix OS.

Re:Debian GNU/kFreeBSD

[Pharmboy]

To add: Yahoo runs their entire infrastructure on FreeBSD, including their ecommerce sites (we have a few), but I haven't seen much in the way of contributing back. On the other hand, Google runs Linux and, well, you know.

Re:Debian GNU/kFreeBSD

[gmack]

Google actually contributes a fair amount back although things could be (and are getting) better.

If you monitor Linux-kernel you would know that.

Re:Debian GNU/kFreeBSD

[Rufty]

I've heard of zombie processes, but a zombie kernel???

Re:Debian GNU/kFreeBSD

[Rufty]

The community is what made linux. Not the license. That's why it's linux *not* GNU/Linux. The community is what matters, not things like gcc, libc or the utils. Those are replaceable, the community isn't.

Re:Debian GNU/kFreeBSD

[Pharmboy]

I did know. I was contrasting Yahoo to Google. Everyone kind of knows that Google gives some back, even via Android by keeping it open. That was the point.

Re:Debian GNU/kFreeBSD

[Rennt]

Linux exists in a near-commodity space where the only difference between it and it's competitors is the license. Competition theory holds a pretty strong argument for causality right there.

Apache, postfix, etc do not compete with Linux - they exist in the same ecosystem.

Re:Debian GNU/kFreeBSD

[Yenya]

That's only partly true. In '94 BSD was even or ahead of Linux in terms of features.

Wrong. As one of the grandparent posts mentions, at that time Linux had shared libraries and compressed kernel, making it much more usable on desktop PCs. I know what I am talking about, I was a 386BSD and NetBSD user then, and for me those features (together with better attitude of the Linux community) were precisely the reasons for migrating to Linux.

Of course, networking at that time was more mature in *BSD. But as for overall system stability, Linux won: around 1994 it was the first UN*X I had that survived running crashme(1) on it.

Re:Debian GNU/kFreeBSD

[Pharmboy]

Suffice it to say that if Yahoo ran Linux, they would NOT be obligated to submit their changes. The GPL puts no restriction on use, only distribution. And "entitlement" is irrelevant to the conversation. You are likely glad you posted AC.

surley OSP

[itzdandy]

spending that cash on a yearly subscription to playboy certainly would have netted the world a new open source porn system providing free as in beer porn to the world!

Re:surley OSP

[Anonymous Coward]

Yeah, but it would have that crappy Finnish porn.

Re:surley OSP

[Anonymous Coward]

Yeah, but it would have that crappy Finnish porn.

You don't care for Tom of Finland? To each his own.

What would the world look like?

[UdoKeir]

That's easy: http://www.guidebookgallery.org/screenshots/win31

Re:What would the world look like?

[ground.zero.612]

That's easy: http://www.guidebookgallery.org/screenshots/win31

Well, that site you posted is suspiciously missing screenies of Trumpet Winsock, and Netscape in the 'internet applications' section. Maybe I was just an 'advanced' user back then...

Re:What would the world look like?

[dunkelfalke]

Nostalgy at its best.

Re:What would the world look like?

[EsJay]

That's easy: http://www.guidebookgallery.org/screenshots/win31

Does anyone else get queasy looking at Windows 3.1?

Re:What would the world look like?

[Hurricane78]

At the end of that link:

Wrong password

This GUI doesn't have this feature.

That cracked me up. XD

Re:What would the world look like?

[BluBrick]

Your Trumpet Winsock and Netscape weren't third party apps like mine? Wow, you really were an advanced user!

Re:What would the world look like?

[cbhacking]

Actually, I was somewhat impressed - compared to modern software it certainly looks dated, but not *as* dated as I expected. Things like the Font dialog, for example, were actually fairly modern in appearance. File Manager looks horribly dated compared to modern filesystem browsers with their previews and so forth, but for its time it was actually very impressive (different icons for file types, graphically hierarchical folder tree, the data at the bottom, and even the MDI - while MDI isn't used much anymore, at the time the ability to do something like look at 4 different parts of your filesystem within one application was pretty cool). Write (replaced by Wordpad in modern versions of Windows) supporting fonts and such on-screen was pretty cool too, even if the interface was lame. Sound Recorder looks shockingly similar to modern Windows versions too. Also, it's amusing how incredibly little Notepad has changed.

Re:What would the world look like?

[ground.zero.612]

Your Trumpet Winsock and Netscape weren't third party apps like mine? Wow, you really were an advanced user!

And? Did you have a point? Because I can see where you tried to go and stopped, and I have a car analogy for you if you really had a point.

Re:What would the world look like?

[BluBrick]

It's not reasonable to expect to see screenshots of Trumpet and Netscape because they were not part of Windows. My point was only that they're not "suspiciously missing", they're just not supposed to be there.

Wait, was that your point in the first place?

Re:What would the world look like?

[ground.zero.612]

It's not reasonable to expect to see screenshots of Trumpet and Netscape because they were not part of Windows. My point was only that they're not "suspiciously missing", they're just not supposed to be there.

Wait, was that your point in the first place?

My point was what the world did look like back then.

The world would be almost the same

[houghi]

We would just have the argue if it should be called GNU/Windows or Windows.

Nobody has said this yet?

[kraemate]

Happy birthday Linus!

Re:Nobody has said this yet?

[PixetaledPikachu]

Happy birthday Linus!

I believe it's on the subject....

Re:happy birthday

[chill]

Its your mom's birthday, too?

Happy Birthday Linux

[mschoolbus]

I, for one, appreciate everything you have done for the computing world!

I hope you have a great birthday and had a merry Christmas!

Alternate timeline...

[Vexler]

If he had bought a Trash-80, would we all be programming Motorola chips today?

Re:Alternate timeline...

[Angst Badger]

If he had bought a Trash-80, would we all be programming Motorola chips today?

Maybe, but it wouldn't be because of the TRS-80, which had a Zilog Z80 CPU.

Re:Alternate timeline...

[freddieb]

Don't you mean Zilog? ie Z80

Re:Alternate timeline...

[ghard]

If he had bought a Trash-80, would we all be programming Motorola chips today?

Maybe, but it wouldn't be because of the TRS-80, which had a Zilog Z80 CPU.

I beg to differ - I've written C on TRS-80 model 16 with XENIX - it had MC68k and Z80 for bootstrap, which was used as I/O co-processor once the 68k system was up, It also had dual 8" floppies, and an 8MB (!!) hard disk which was larger than a modern desktop PC chassis.

Also the parent got the time wrong - affordable i386-based systems didn't really co-exist with TRS-80 - and the price of a Trash - even a Z80-based model would've definitely been more than you'd be able to buy with your Xmas cash in Finland those days.

interception Osama ICE freedom Abbas COSCO threat morse Comirex AMEMB
Peking Defcon DES INS Attorney General

Re:Alternate timeline...

[sznupi]

Actually, Amiga would have been also a very valid choice back then, at least in Europe. I wonder if Linus ever said why he went with a PC.

Re:Alternate timeline...

[ghard]

Actually, Amiga would have been also a very valid choice back then, at least in Europe. I wonder if Linus ever said why he went with a PC.

Lack of MMU, perhaps?

Mena Baranyi Verisign embassy anarchy domestic disruption national
information infrastructure diwn [Hello to all my friends and fans in
domestic surveillance] Jyllandsposten Sundevil UNSCOM Qaddafi Consul
Arnett

Re:Alternate timeline...

[hitmark]

was the TRS-80 even available outside of USA?

btw, he started out on this:
http://en.wikipedia.org/wiki/Sinclair_QL

if the opening of the article is correct. Only when he reached university did he get the 386...

Re:Alternate timeline...

[sznupi]

Not in case of A3000, which would be likely equivalent. Or A2000 with upgraded CPU.

Re:Alternate timeline...

[Angst Badger]

I yield the point to your more extensive knowledge. My experience with the TRS-80 stops with the Model III, after which my school brought in PCs, 386-based, I think. All kinds of anachronisms appear in cash-strapped schools: my favorite machines in the lab were a pair of IBM System/3 Model 10 minicomputers, refrigerator-sized beasts with dials and blinkenlights and card readers which, by the time I parted ways with them in 1988, had been out of production for more than twelve years. But how else would I have reimplemented troff in RPG-II? (The other options were COBOL and FORTRAN-IV.)

Re:Alternate timeline...

[ghard]

Our school (in Finland) had a TRS-80 model I, on which I learned my first assembly language programming skillz. My dad had a model III. I myself did some contracting work when 15yrs old on a model 16, which was actually an I/O / modem bank front-end for an order processing system on an IBM system 38, with mobile clients using the model 100, sending the orders in and getting their itineraries back using slightly modified XMODEM-protocol.

So the short answer. Yes, they were available outside of USA and not just as the Dragons (which were rebranded Tandy/TRS-80 color computers.)

blinded

[prgrmr]

He could have bought an Official Red Ryder, Carbine-Action, Two-Hundred-Shot, Range Model Air Rifle.

And then shot his eye out.

Re:blinded

[mcgrew]

At age 19? Hardly! You'd have to be pretty stupid (or six years old) to do that.

BTW, I hated that corny movie.

Re:blinded

[fdrebin]

I must respectfully disagree. I happen to be an airgun nut, and have airguns in my collection that run in the $500 range - and there are plenty that cost a lot more. I don't have a a Red Ryder because I collect shooters, not collectors items, but there are plenty of folks who would and do purchase such items as the 'original' and modern replicas.

/F

this is what happened

[Device666]

If linux wouldn't bought a computer, this year would be probably the year GNU\Hurd would be finished. Gnome KDE etc would probably not exist.

Re:this is what happened

[Anonymous Coward]

His name is Linus. Your grammar is awful. Your nick name is 12 year old'ish.

Re:this is what happened

[LOLLinux]

If linux wouldn't bought a computer, this year would be probably the year GNU\Hurd would be finished.

Based on what reasoning? Hurd is the Duke Nukem Forever of OS kernels. It would have self-imploded with or without Linus and the Linux kernel.

Re:this is what happened

[Device666]

I am not a native American speaker, so I am allowed to mutilate your language like there is no tomorrow. I texted my message from a very small hightec handheld device made in America, so in some lightly indirect way the blame is on yourself. It's an evil device. I thought of Linux when trying to type Linus, even Linus himself has that kind of typos. So it's his fault also to give Linux a too similar name. Device666 isn't my nickname. My parents gave me that name. But I guess Anonymous Coward is not your nickname either?? And yes, you are very funny! Mod +1

We will never know

[prefec2]

Maybe without Linux we would use Minix or Hurd today. While Linus caused an crystallization point for hundreds of developers he did not write the thing alone. these people were already there. More or less waiting for something like this to happen. Most of them were already part of the Minix mailing list. So most likely Linux was already waiting to happen then. From my own time as an undergraduate. all the good programmers wanted to write an OS. And when it Linux came into existence everyone said cool. I take it and I do something with it. The same happened later with the browser as well. And if X11 would have had a better programming interface there would have been more different browsers out there. Still. Thanks to Linus for starting it.

Re:We will never know

[hitmark]

minix was in no way free at the tho, iirc. It only became free in response to the success of linux...

hell, thinking about it, i wonder if the monolithic vs micro-kernel debate was a indirect case of fud marketing for minix...

Re:We will never know

[Kjella]

While Linus caused an crystallization point for hundreds of developers he did not write the thing alone. these people were already there. More or less waiting for something like this to happen. Most of them were already part of the Minix mailing list. So most likely Linux was already waiting to happen then. From my own time as an undergraduate. all the good programmers wanted to write an OS.

A lot of people are in the wishful thinking brigade of "It would be cool to..." without ever walking the walk, they just like to dream about it. I know it with myself that it's pretty easy to dream up grand projects, something completely different to drive them to practical completion. I'm sure some people would have created a basic OS just to let it fizzle for lack of interest or because of final exams or because they got a job or got a girlfriend or family or didn't really like all the hard real-life problems or didn't cope well QA'ing code submissions as opposed to coding or any one of the million things that leads to dead halfassed projects on sourceforge. Linus wasn't just the spark that set off a community, he's been the one driving the community with massive amounts of code written by him personally and the persistance to drive the project to maturity. There's a lot of people who could have played the supporting roles, but there's few who could have walked in his shoes. I'm not so sure anyone would have.

Re:We will never know

[dgr73]

I thank Tommy the Clown... err.. Linus Torvalds for starting the movement, but what we do now is completely different... riiiighht..

Re:We will never know

[captaindomon]

Blasphemer! None shall belittle His Linuxship! He suggesteth that Linus did not create the world in seven days, but instead the Linux world evolved over time, from earlier ancestors, with many different contributions! Burn the witch!

Re:We will never know

[LWATCDR]

I would bet on BSD. BSD was already around and once the court case was finished BSD would have jumped to the front of the pack IMHO.
Most of the GNU software would have been ported very quickly and KDE and probably GTK would have been written for BSD.
Hey I am a Linux user but the simple truth is if had not been for Linus it really would have been somebody else. The 386 and GCC was available as was Tanenbaum's book.
I think that Tanenbaum gets a little short changed in the whole free OS world. No he didn't write Linux and Linus didn't copy Minix. What did happen is Linus and probably every other person that worked on Linux back then did was read Tanenbaum's text book. In that book he showed that it was possible for a human to write a real OS and even showed them how to do it.
Tanenbaum did what he wanted to do. He taught a lot of people how too write an OS.

Alternate Reality?

[The Redster!]

2010 will be the year of Solaris on the Desktop!

DX or SX?

[anagama]

Wasn't it the SX line that had the math coprocessor disabled? My first computer (not counting my CoCo) was a 386 SX 20, which was cheaper than a DX for want of a coprocessor.

Re:DX or SX?

[mother_reincarnated]

That's what I thought too, but we're just confusing it with the 486DX/SX... The 386SX had a 16bit external bus.

Re:DX or SX?

[Stonent1]

The 386SX was a 32 bit processor internally but had a 16 bit data bus. The 386DX was a straight 32bit processor all the way through. There was a third flawed varient that had a problem switching between real and protected mode that could lock up the system. Those chips would be stamped that they were only certified for 16 bit apps. The ones that tested good had a double sigma stamp on them. Neither the 386sx nor 386dx had math coprocessors. The 486 however was a different story. The 486DX had a coprocessor but the 486SX did not.

Re:DX or SX?

[dantum_sh]

I'm pretty sure DX is the model with coprocessor. The SX were the crippled chips that didn't have coprocessor.

Re:DX or SX?

[blind biker]

The 386SX was a 32 bit processor internally but had a 16 bit data bus. The 386DX was a straight 32bit processor all the way through. There was a third flawed varient that had a problem switching between real and protected mode that could lock up the system. Those chips would be stamped that they were only certified for 16 bit apps. The ones that tested good had a double sigma stamp on them. Neither the 386sx nor 386dx had math coprocessors. The 486 however was a different story. The 486DX had a coprocessor but the 486SX did not.

And let's not relent in confusing the audience: both 486SX and 486DX had a 32 bit data bus :) And then came the 486DX2 which ran at 66 MHz instead of the 486DX's 33. The 486DX4, unlike what you'd expect, didn't have a 4x33 MHz clock speed, but a 100 MHz one. If you're not confused by now, you're not human.

Re:DX or SX?

[martinde]

What's really funny is you could buy a 487 coprocessor if you had a 486sx:
http://en.wikipedia.org/wiki/Intel_80487#80487

I.e. you could plug in the chip you REALLY wanted into a second socket.

Re:DX or SX?

[d36]

than what am I?

Re:DX or SX?

[Rufty]

Generally, what happened was 486DXs were made. If there was a flaw in the maths unit, that got disabled and it got badged 486SX. Flaw in the main CPU, but maths unit OK? Stand alone coprocessor. Same silicon.

Doesn't qualify for one-name status

[sunderland56]

Mr. Torvalds may be well known, but when you use just the single name "Linus", most people think of the blanket-carrying kid in Peanuts.

Re:Doesn't qualify for one-name status

[Anonymous Coward]

but when you use just the single name "Linus", most people think of the blanket-carrying kid in Peanuts.

That's a hell of a blanket statement for you to make.

Re:Doesn't qualify for one-name status

[Dunbal]

most people think of the blanket-carrying kid in Peanuts.

      Perhaps anywhere else, but not here on slashdot. And a heads up: RMS usually refers to Richard Matthew Stallman, not Root Mean Square... even though most of us here know the uses of the latter.

Re:Doesn't qualify for one-name status

[mcgrew]

Didn't you know? The blanket-carrying kid grew up and wrote an operating system kernel!

Re:Doesn't qualify for one-name status

[British]

This is slashdot. Half the acronyms posted to the news headers are lost on me.

Re:Doesn't qualify for one-name status

[SpooForBrains]

Not if you pronounce it correctly they don't.

Re:Doesn't qualify for one-name status

[AvitarX]

On slashdot?

Re:Doesn't qualify for one-name status

[Thumper_SVX]

most people think of the blanket-carrying kid in Peanuts.

      Perhaps anywhere else, but not here on slashdot. And a heads up: RMS usually refers to Richard Matthew Stallman, not Root Mean Square... even though most of us here know the uses of the latter.

And if you hear the denizens referring to a RIM job, it's getting employment at the maker of Blackberry...

Re:Doesn't qualify for one-name status

[larry bagina]

Linus (Peanuts): Sounds like "Vaginas"
Linus (Torvalds): Sounds like "Penis"

What if Tanenbaum hadn't written MINIX?

[munitor]

That's the real question. But today is about Happy Birthday to Linus!

386dx, no coprocessor?

[damn_registrars]

Perhaps my memory is incorrect, but I thought the difference between the 80386sx and 80386dx was that the dx had a built-in math coprocessor.

Re:386dx, no coprocessor?

[UdoKeir]

No, that's the 486dx/sx you're thinking of. The 386sx had a 16-bit data bus, half that of the dx.

Re:386dx, no coprocessor?

[gregarican]

Nope that distinction didn't find its place until the Intel 80486 line of CPU's. Back during the 80386 days it was only differntiating between 16 and 32 bit handling --> http://encyclopedia.thefreedictionary.com/386+DX.

Re:386dx, no coprocessor?

[dee.cz]

DX had wider bus, not a coprocessor. Must have been insanely expensive when Linus bought it, I switched to the same setup 16.5 years ago.

Re:no co-proc?

[jgardia]

386SX -> no math coprocessor, 16bit data bus 386DX -> no math coprocessor, 32bit data bus 486SX -> no math coprocessor, 32bit data bus 486DX -> math coprocessor, 32bit data bus

Happy Birthday!

[dwiget001]

And, many happy more!

Re:Happy Birthday!

[Lord Lode]

Yes, Linus, if you're reading this, happy birthday! :)

Birthday money

[Megaweapon]

Perhaps better spent on a $699 license from SCO. /sarc

Re:DX33

[dingen]

No, that's the 80486. With the 80386, DX meant you got a full 32 bit CPU, instead of partly 16 bit one with the SX.

no, you're wrong

[OrangeTide]

You fail your nerd lore. Turn in your nerd badge immediately.

80386DX had a 32-bit bus and came in a small PGA package. and 80386SX was only 16-bit(24-bit address) and generally was not socketable. There was a mathco for each, the 80387DX was a PGA package, the SX was a PLCC package.

80486DX has a math co and the SX did not (both 486 models were 32-bit bus and could fit in the same socket)

I miss my 80386DX+80387mathco system. it was a sweet setup.

Re:no, you're wrong

[shentino]

Indeed. The 386 SX was designed as a socket compatible replacement for the 286

Re:no, you're wrong

[OrangeTide]

no, it was never socket compatible with the 80286. The 80286s were typically PGA or PLCC, while 80386SX were only a QFP as far as I know. although most of the support ICs for the 80286 could be used on the 80386SX making them cheaper to manufacture. but 16-bit bus motherboards were in general cheaper to manufacture.

It's a little irritating that you state something as fact that can be shown to be false by cracking up a book or search engine.

Re:no, you're wrong

[shentino]

The 386SX was still designed with 286 compatibility in mind. 24-bit address bus and 16-bit data bus.

Even if I didn't nail it I was still not far off.

Most Americans might

[Kupfernigk]

Some North Americans really do overestimate the penetration of their own popular culture. Globally, most people wouldn't think of anyone at all. On Slashdot - Mr. Torvalds.

Re:Most Americans might

[Eevee]

And some people underestimate it. According to wikipedia:

At its peak, Peanuts ran in over 2,600 newspapers, with a readership of 355 million in 75 countries, and was translated into 21 languages.

That's a bit more than just North America.

Re:Most Americans might

[amorsen]

Names were translated too though.

Re:Most Americans might

[weicco]

Here in Finland Linus van Pelt = Eppu Peltonen

That's my first computer too

[dingen]

DX33 80386, with 4 Megs of RAM, no co-processor, and a 40 Megabyte hard disc

That's almost exactly my first computer too. Altough I really had a 20 MB harddisk, but I used doublespace to get 40 MB. And I didn't have the Intel DX33, but the Cyrix DX40 instead. That 7 MHz really made the difference.

Re:That's my first computer too

[XanC]

There wasn't an Intel 40MHz 386.

Re:That's my first computer too

[trash eighty]

My first computer (that i bought myself) was a Unisys 386SX-33, 4MB of RAM. i had a massive 80MB harddrive though. for some reason i decided to install OS/2 2.1 on it.

Re:That's my first computer too

[dingen]

I have a box of OS/2 Warp 3 on a shelf, so I suppose I've run that as well at some point, but I really can't remember anything about it except for staring at the sand clock pointer forever :P

Re:That's my first computer too

[dingen]

Exactly. Both Cyrix and AMD offered 386's and 486's at higher clockspeeds and for less money than what Intel was selling. I also had a 80486 DX later which ran at a crazy 125 MHz. It wouldn't be for much, much later that Intel sold chips at those frequencies.

But anyway, the whole game changed when the Pentium came around and Intel could actually patent their chip for the first time. Cyrix was dead a short time after and AMD was forced into the niche were it remained till this very day.

Re:That's my first computer too

[dingen]

I'm 100% sure it was a 80386 DX at 40 MHz. Not by Intel though. I thought it was a Cyrix chip, but it could've been AMD as well.

Re:That's my first computer too

[An ominous Cow art]

I don't remember whether Cyrix made a 40Mhz 386 or not. My first IBM PC-contemptible system was an AMD 386-40, 16M of RAM, with an Adaptec 1542B & 212M SCSI disk, a Diamond Stealth video card (I think; that might have come later), Sound Blaster, and combo 3.5"/5.25" floppy. A superb machine for the time, but I think I paid $2000 for it.

Re:That's my first computer too

[dingen]

I remember $2000 being a very middle-of-the-road and maybe even modest amount to pay for a new computer system back in the late eighties and late nineties.

Re:That's my first computer too

[aztracker1]

My first was an IBM PC-XT with an 8mhz (iirc) 8088 cpu, 512KB of ram and a whopping 10MB MFM Full Height 5.25" HDD (full height is twice what a 5.25" bay is today). My second was an AMD 5x86 133Mhz cpu with 256MB of ram and a 160MB HDD. The XT's hard drive ran so slow, it was actually faster with stacker. I upped to 640KB, EGA and a 9600kbps modem before moving on to something faster. I think if I had to go back to it today, I'd probably shoot the thing.

Big deal.

[jgreco]

No Linux? Big deal. There are plenty of alternatives. Perhaps we would have ended up without a GPL-encumbered OS - maybe something like NetBSD or FreeBSD - doing these things. Nothing to see here.

No coprocessor...

[TrisexualPuppy]

The point about Linus's machine having no coprocessor is actually true. This made development a little iffy since floating point math had to be done in software. The i386DX actually did not have an FPU, and the coprocessor was the i387 which was not all that popular but was compatible with both the DX and SX models. It was not until the introduction of the i486 that the SX had no coprocessor and the DX had a built-in coprocessor.

Interesting read.

Re:No coprocessor...

[Anonymous Coward]

The 486 SX was just a 486 that they could not guarantee that the coprocessor would work so it was switched off.

Re:No coprocessor...

[vegiVamp]

Yep. I remember buying one of the things. Was the first time I opened up the magic box myself, too :-)

Re:No coprocessor...

[teslar]

The point about Linus's machine having no coprocessor is actually true

Nineteen years ago or so, I also got my hands on my first PC, pretty much same specs but with the coprocessor. My programming achievements at the time were pretty much limited to batch files. Linus wrote an OS on smaller hardware. Kinda makes me feel like I wasn't really using the full potential available to me :)

Of course, I was 10 years younger at the time :)

Re:No coprocessor...

[Sauron23]

The 80387 was the first chip I installed and it was in a 386SX. I wasn't trying to change the world with an open kernel or even learn to program. It was Doom. I had to get Doom running. Oh no! Hammer pants and wine coolers! The Cure! I think I better go lay down now.

Re:No coprocessor...

[Godji]

I''ll go for a wild guess here, but I bet you're still 10 years younger :)

Re:No coprocessor...

[dotgain]

That's the first time I've ever heard this, but I'll believe it. Nothing, I repeat nothing, will ever surprise me about the evolution of x86 any more.

If you were to tell me that Little Endianness was simply the result of someone putting something on an overhead projector the wrong way, I'd believe you (because it seems like an extremely fucking stupid idea otherwise: "2 ^ 16 equals five-hundred-thirty-six, sixty-five thousand"

If you were to tell me that the Pentium was really 64-bit, but the fabricators never hooked up the address pins because they never got the memo, I'd believe you.

No doubt, x86 is the cheapest, fastest and most prevalent CPU in computers today, and probably always will be, but fuck me if it doesn't look like the biggest kluge in the world.

Re:No coprocessor...

[Urkki]

That's the first time I've ever heard this, but I'll believe it. Nothing, I repeat nothing, will ever surprise me about the evolution of x86 any more.

I seem to remember that this was the situation originally. Not that surprising really, when you think about it. FPU was a sizable chunk of the processor die (caches were still very small those days, even though 486 already had L1 cache) and being able to sell the processors that had faulty FPU must have increased yields quite nicely. Later on they made pure 486SX processors that didn't have FPU on them, I believe.

All from memory, could be totally wrong, read above at your own risk..

Re:No coprocessor...

[shentino]

Coprocessors were rare enough that it was not a good idea to have a "do or die" dependency on them, so it paid to be good with emulation.

Re:No coprocessor...

[matt_hs]

The 486 SX was just a 486 that they could not guarantee that the coprocessor would work so it was switched off.

This was how I remember hearing it as well. In addition, the 487SX was a 486 that couldn't guarantee the main processor so the main processor was switched off. I believe i486 boards used to have a co-processor slot that could take either an 80387 or an i487SX.

Re:No coprocessor...

[yuhong]

On the matter of the coprocessor, here is a newsgroup post by Linus asking about the coprocessor before the release of Linux 0.01:
http://groups.google.com/group/comp.sys.intel/browse_frm/thread/c0b4d0b80527cfba/a6d2cc19bfb3ba1f
Anyone on slashdot old enough to answer this one?

Re:No coprocessor...

[ls671]

> The 486 SX was just a 486 that they could not guarantee that the coprocessor would work so it was switched off.

Exactly, I thought mostly every /. reader knew that but reading the comments and replies makes me feel a little old... ;-)

In short SX and DX where made at once, then on the testing stage, if the co-processor failed, they sold it as a SX, if it worked well, they sold it as a DX.

Note that this principle is still applied today, I wrote about it previously to explain why overclocking sometimes work and sometimes don't:

http://slashdot.org/comments.pl?sid=1246015&cid=28105441&art_pos=5

Re:No coprocessor...

[turgid]

This was how I remember hearing it as well. In addition, the 487SX was a 486 that couldn't guarantee the main processor so the main processor was switched off. I believe i486 boards used to have a co-processor slot that could take either an 80387 or an i487SX.

Not quite.

The 487SX was a full 486DX but sold as a coprocessor (a little cheaper than a full 80486DX). You could (and people did) buy a 486 motherboard and an 80487SX only instead of the more expensive 80486DX.

The 80487SX was a triumph of intel marketing over reality.

The 80486 had a build in 8k (IIRC) L1 cache as well as FPU. The 80386 had no cache. The 80486 was optimised such that many of the common instructions ran faster that the 80386 (and similarly for the 80387). Back in the day, with unoptimised software (i.e. compiled for 286) the 486 was about twice as fast as the 386 at the same clock speed. Compared to a 386 without a 387 copro (which was identical to one of the 287s incidentally), a full-blown 486DX was an order of magnitude more powerful. Doom anyone?

Re:No coprocessor...

[rickb928]

Someone else by now must have pointed out that a DX33 80386 is probably not what Linus bought.

He probably bought a 486-DX33.

I know that the first version of Linux I ever ran I installed on my 486-DX25, with 4MB RAM and two 40MB SCSI drives, which was running NetWare 3.01 at the time. what a bear, Slackware 0.9 was a struggle but I got it out of a book called 'the Internet CD' by Vivian Veou. What an awesome book!

Then I got a 486-DX2/66 and another 4MB RAM. And some bigger drives. And I 'demoted' the DX25 to 'just' Linux.

For a while I was running MySQL, PHP, Apache, and Perl on both NetWare and Linux. Was fun. Making Squid work like BorderManager was impossible, but Squid actually did what I wanted. BorderManager did enhance my Internet experience, mostly by blocking doubleclick and many other ad sites, back when that was easy. But I had my own email server out there running either Red Hat or Fedora for over 12 years, probably more. Now it's hosted, and I miss updating it and fussing over the hosts.deny and Sendmail settings. Well, not that much, I guess.

After about NetWare 5.x, the thrill was gone. But Linux is not. Thanks, Linus.

Re:No coprocessor...

[yuhong]

And the fact that Linus's computer initially had no coprocessor probably made it harder.

Re:No coprocessor...

[quenda]

coprocessor ...was switched off.

Luxury!!
My first CPU was the Intel 8080. That did not even have integer arithmetic!
We had to learn how to do our own binary long-division and multiplication using shifts, adds and subtracts.
Nothing but 1s and 0s. And sometimes we didn't even have 1s.
Tell that to young people today ....

Re:No coprocessor...

[someone1234]

Yeah. Little endian makes much more sense "inside" the computer. Even Apple converted to little endian not so long ago.

Re:No coprocessor...

[Just Some Guy]

If you were to tell me that Little Endianness was simply the result of someone putting something on an overhead projector the wrong way, I'd believe you (because it seems like an extremely fucking stupid idea otherwise: "2 ^ 16 equals five-hundred-thirty-six, sixty-five thousand"

Here's the logic, as explained to me:

Assume a 32-bit word starts off zeroed. If you store an 8-bit value at the first byte in the word, then interpret that whole word as a 32-bit value, then you'd get the same answer back. For example:

Starting: 0x00 0x00 0x00 0x00

Store 0x01: 0x01 0x00 0x00 0x00

As a little-endian 32-bit value: 0x00000001

As a big-endian 32-bit value: 0x01000000

Taken in the context of an era when people used a lot of 8-bit values, casting them upward to 16 bits as necessary (which was quite often), it almost kinda makes sense.

Re:No coprocessor...

[ZaphDingbat]

If you were to tell me that Little Endianness was simply the result of someone putting something on an overhead projector the wrong way, I'd believe you (because it seems like an extremely fucking stupid idea otherwise: "2 ^ 16 equals five-hundred-thirty-six, sixty-five thousand"

Little endianness allows you to do type casts without adjusting pointers.

Re:No coprocessor...

[Sauron23]

I figured that out immediately. Which was why I thought it was funny. My Doom install consisted of a stack of black floppies labeled D1, D2 ..

Re:No coprocessor...

[BlindBear]

I had a Z-80 for my first effort, it had some 1s , but not many ... but when you are flying at 1.77Mhz you don't need that many of them.

A undisclosed person said at his party

[jhoegl]

"What a nerd!"

Time travel

[mariox19]

Everyone kills Hitler on his first time travel... and then there's Bill Gates.

Re:no co-proc?

[Jeremy Erwin]

A DX33 had a math co-processor,
Only if you also installed a 387. Or perhaps a Weitek Abacus 3167.

Re:get real

[DrgnDancer]

Whether he's a douche or not is immaterial to his accomplishments. He wrote (originally) and later managed the development of the kernel that allowed GNU to become part of something mainstream. Whatever Hurd may have done, the remains that it was dramatically behind schedule and short of developers when Linus produced his first kernel; and didn't manage to catch up in either capability or developer interest during the years that Linux was little more than a barely functional hobby system. Douche or not, Linus managed to develop a system good enough to interest people, then hold onto their attention long enough to turn it into something useful.

RMS not an egomaniac? Seriously? He's the worst kind of egomaniac, the "Look at me I'm humble! Look! Me! Humble! Just wanna help, but can't help unless you. Look. At. Me!" kind. Flagellating yourself in the public square is not a gesture of humility. None of which should take away from his accomplishments either, don't get me wrong. Just because I happen to think he's a douche doesn't mean he's not a very effective douche, just like your opinion of Linus doesn't change how effective he's been.

As long as we're asking what if's...

[dclozier]

What if Watson was right? :D

I think there is a world market for maybe five computers (quote may or may not be his)

Re:As long as we're asking what if's...

[Sir_Lewk]

What if Columbus was wrong? At a certian point, such speculation becomes meaningless ;)

Re:As long as we're asking what if's...

[ja]

What if Watson was right? :D

How many did he actually sell?

My best guess

[Zero__Kelvin]

I'm an optimist, so I'd like to think it would look something like this.

Powerful computer

[Dan East]

For his age, that was a pretty powerful first computer. I'm a few years younger than Linus, and my first computer was a TI-99/4A, followed by an Amiga 1000 (512K RAM, no HDD). I think many people of our generation started with floppy-based computers (Apple II, TRS-80, VIC-20, C64, Amiga) with less than 1 MB RAM. I saved up for and purchased the Amiga from my job as a bagger at a grocery store. Paid $750 for it used, and it came with a monitor and an external floppy drive (really saved on the disc swapping having two drives!). That was right when the Amiga 500 was released.

Re:Powerful computer

[drinkypoo]

My first computer was an Amiga 500, for which I paid around $600 with the RF adapter... and I hooked it right up to my TV and learned to live with badly flickering interlaced graphics. And I only had one floppy drive, oh the humanity. Actually I had a C= 16 before that, but no storage device so it hardly counts. But lots of my peers started out with cassette storage...

Re:Powerful computer

[Amouth]

http://en.wikipedia.org/wiki/Amstrad_PCW

we had one of the first ones - i remember having to trouble shoot it once - turns out it couldn't boot unless the printer was attached.

Re:Powerful computer

[aztracker1]

Agreed, I could barely afford a used XT at around the same time... Probably why I now have all Unicomp keyboards.

What kind of a Pentium is that?

[Tibor the Hun]

Heey, DX33? What kind of a Pentium is that?

Grattis på födelsedagen!

[CaptainOfSpray]

Ja må han leva uti hundrade år!

OK, I'm showing off - I lived in Sweden 18 years, became fluent in Swedish, and I'm guessing (from his name) that Linus is mother tongue Swedish rather than Finnish.
But we're raising a glass and shouting "Skål" and "Gippis" and so on...

Re:Grattis på födelsedagen!

[ja]

... och sedan ska han skjutas, i en skottkärra fram!

Re:Grattis på födelsedagen!

[Anonymous Coward]

For you English-only speakers:

och sedan ska han skjutas, i en skottkärra fram!

translates to:

Eight cars can have skates, and in scottcar (we can finish before the others get there).

It is a corruption of an old Swedish proverb that says that because your carriage has all the glitter and trappings, doesn't mean it will work all that well. I think the OP was referrring to Windows versus Linux.

Re:Grattis på födelsedagen!

[Katchu]

Ei, se on: Hyvää syntymäpäivää, Linus.

Re:Grattis på födelsedagen!

[Culture20]

Finnish wins. Six umlauts and a lower case "i", that's 13 dots!

Re:Grattis på födelsedagen!

[21mhz]

That's because of the vowel harmony: if a (simple) word has front vowels, that is ä, ö, or y (pity it's not ü), this means all vowels have to be front or neutral. So words tend to be either umlaut-rich or without umlauts at all.

But really, Linus's native language is Swedish, even though he most likely knew Finnish very well.

Re:Grattis på födelsedagen!

[master_p]

I wonder why I remembered about this...

Re:Grattis på födelsedagen!

[V!NCENT]

http://www.youtube.com/watch?v=RYQUsp-jxDQ

It sounds all the same... :')

Re:Grattis på födelsedagen!

[ja]

I seem to recollect having ...

no idea whatsoever what you are talking about?

Re:Grattis på födelsedagen!

[Rennt]

Hey, I'm just impressed you got slashdot to display those characters properly, show off!

Re:Grattis på födelsedagen!

[CaptainOfSpray]

Hoho, I cheated of course. Typed them into Notepad, then cut-and-pasted.

Re:Grattis på födelsedagen!

[ja]

OK, we are slowly getting there. Still missing the part where "to be shot" is spelled the same way as "to be pushed (in a wheelbarrow)"

Re:get real

[LOLLinux]

Open source would be a lot better off with a few less egomaniacs like Linus and a few more - dare I say it? - RMS's.

RMS not an egomaniac? lolwut? The whole GNU/Linux or GNU+Linux thing is nothing but pure egomania on RMS's part.

I'm amused by all the HURD references

[afabbro]

The open source "competition" to Linux has been *BSD. If Linux had never existed, we'd all be running *BSD. End of story, really. And it would have happened quickly - if memory serves, the only reason Linux took off was because BSD was still in or had just gotten out of the long clusterfrack legal disputes. If there had been no Linux, *BSD would have picked up its steam, only a year or two later.

At some point, someone would have married the best parts of GNU with *BSD and you'd have RMS screaming about GNU/BSD.

HURD, for all practical purposes, has never existed.

Re:I'm amused by all the HURD references

[scharkalvin]

While the use of BSD code in a GPL'd project is possible (because the BSD license lets you do ANYTHING!) I'm not sure the inverse is true. So if the BSD kernel became the mainstream instead of Linux the only way it could use GNU/GPL code would have been to wrap the entire project in GPL. I guess this is what Debian is doing. However BSD has it's own versions of the GNU programs and stands quite well on it's own.

Re:I'm amused by all the HURD references

[selven]

You just have to separate out the GPL bits from the other bits. Mac OSX includes quite a lot of GPL software out of the box.

Re:I'm amused by all the HURD references

[aztracker1]

You can run GPL'd device drivers, and closed source software under BSD just fine. It's actually more permissive as an underlying OS than Linux is. Also, traditionally BSD has had much better binary compatibility, though the past 5 years or so there have been a lot of major version changes, with some breaks, still better than graphics driver compatibility in windows.

Re:I'm amused by all the HURD references

[spitzak]

Userspace programs can be GPL with no problem and the kernel and other programs can still be BSD. Considering that is exactly what is happening today on BSD it should be obvious there would have been no problem with "GNU/BSD" as RMS would want to call it.

What if Linus had bought a mac?

[Ukab the Great]

What would the world look like had he spent his money on something else?

Maybe he would have bought a mac, developed an appreciation for user experience design at the start of a project, collaborated with usability experts to design a free standardized user friendly UI when he first started work on Linux, and today Linux on the desktop might be light-years ahead of where it currently is.

Re:over 40

[rolfwind]

Is he going to invent anything else? Or after 40 do you just give up on life?

http://en.wikipedia.org/wiki/Linus_Torvalds#Authority_on_Linux
"About 2% of the Linux kernel as of 2006 was written by Torvalds himself.[13] Since Linux has had thousands of contributors, such a percentage represents a significant personal contribution to the overall amount of code. Torvalds remains the ultimate authority on what new code is incorporated into the standard Linux kernel.[18]"

Do you know how much output that is?! Also, consider for a minute, that Linux isn't like the lightbulb, invent once and the work is done. How far would linux have gone if work quit in 1991, 1995, 2000? It's a work-in-progress.

The world is littered with half-assed and half-finished projects, particularly software. It's far better that Linus brings and continues one project to excellence than do a dozen mediocre projects that quite never get there.

Maybe you should go out and invent something. If it had 1/100 of the impact Linux has, you'll the world for the better significantly.

Re:over 40

[mrjohnson]

What about git? It's slowly taking over distributed source control...

Thinking ahead

[WaroDaBeast]

When Linux is installed on every computer in half a millennium time, we'll be living in a perfect, brave GNU world. ;-)

NeXT

[Darth Cider]

I wanted a NeXT at that time. Man, $6500! But there was no Photoshop equivalent for NeXT, despite their photoshopped brochures, so I called their office in California (seriously) to see if they had any image manipulation software. The person on the phone, a very nice woman, only had scripts to read from. Later that week, though, I happened to see a piece of mail sent from the Free Software Foundation to a professor at my university. (Just the return address, not the contents of the letter.) That's when it first hit ME, that collaboration was an unstoppable idea, because code is modular and such. I had a terrible notion that it would all happen really quickly, that if I borrowed money to buy a NeXT, that free Photoshop-equivalent software would be available almost right away. I'm glad I waited for OS X.

So it's good that Linus came along with a much more realistic idea of time and what could be done with it, with perseverance. Kudos and happy birthday, Penguin person.

Re:over 40

[jo42]

Everyone knows "42" is the real milestone.

Re:get real

[unixfan]

I agree with you. Though I really doubt all the different Unix like O/S's would have grown anywhere near what it has if Linux had not created all that interest w Linux. They had been around for decades but never managed to generate anything like the following Linux did. All the different BSD's got a huge boost from the attention Linux generated. I seriously doubt any of them would have grown much by themselves. Linux had, obviously, just the right mix to take off as it did.

Oh, and Happy birthday Linus! :)

Props to Linus

[StuartHankins]

Thanks for all the hard work you've done and the world you've helped open! Enjoy your birthday!

If he spent his money on something else?

[okmijnuhb]

What would the world look like had he spent his money on something else?

Well maybe the world's computers would run on hookers and blow?

Other things could have happened

[Stregano]

Like if he spent his money on porn

Re:Other things could have happened

[selven]

GNU/porn?

Re:over 40

[Antique Geekmeister]

Linus invented 'git' much more recently, in 2005. If you haven't reviewed it for source control, and compared it to Subversion at Subversion's expense, I urge you to do so. It is lighter weight, _far_ faster, allows remote development far more easily, and actually pays attention to security with its far better handling of SSH keys and its built-in GPG signatures for software tags.

I can also attest that you only give up on life at 40 if your first 40 years weren't worth living. And in that case, your age probably wasn't the problem.

What if

[MoZ-RedShirt]

What would the world look like had he spent his money on something else?

Good question ...
http://www.youtube.com/watch?v=OGegKKF2Xpw

Re:Actually... He did!

[LaughingCoder]

If Bill Gates contributed to humanity, I think his contribution was pretty negative.

And you'd be dead wrong.

http://en.wikipedia.org/wiki/Bill_&_Melinda_Gates_Foundation

We would probably have Gnu with another kernel.

[miffo.swe]

While i dont want to belittle Linus work i think that GNU would just have another kernel and be about as popular as of now.

The reason i think that is that im my mind the GPL license and GNU was the single most significant factor to GNU/Linux success. If Linus had went for eg. a BSD license Linux would be a mostly one man show were people took Linux code, altered it and didnt return anything back. Now, thanks to the GPL, we have numoreous companies working together on the same code and nobody can take the code and hide it away.

The Linux kernel wouldnt do that much by itself without all the various GPL code that it fits in with. Personally i would love if there was more active kernels in use with GNU. A standardized API for drivers would also be nice. That way it would be immensely easier for new kernel/OS projects to lift off the ground.

Re:We would probably have Gnu with another kernel.

[SwedishPenguin]

I don't know about Apache, but there are several proprietary products based on Eclipse that likely have not contributed any source back to the main project.

Re:We would probably have Gnu with another kernel.

[welsh git]

Just because they can, doesn't mean they WILL. That's the main problem with the GNU fanbois.

Yahoo committed accept-filters back to FreeBSD.

Apple has released code and bug fixes.

There are MANY MANY more examples.

And it doesn't have to be done in good faith. There are many good reasons for a company to contribute code back to the main stream..

If you don't contribute the changes back, then every time you update the OS, you have to repatch your system with your changes, and may even have to do alot of work if there are API changes.

By committing the code to the project, you are reducing the amount of work you'll need to do, as other people will make sure your code remains current.

Re:over 40

[gabba_gabba_hey]

Seconded. git is fantastic. I use it pretty much daily.

Re:over 40

[Godji]

Oh, and by the way, Git.

Sad this is +5 informative

[anti-NAT]

Surely everybody on Slashdot is old enough to remember 386s? What about 286s? I even remember a 186 ... It scared me that where I worked recently with a lot of Gen Y's, some of them may not have ever used a single-tasking OS like MS-DOS or CPM.

Re:Sad this is +5 informative

[mitgib]

Mine was the Atari 400 with a cassette tape drive for storage, then around 83 I bought a TRS80. Wasn't until the early 90's that I bought a PC, a 386SX33 with 2mb of ram and a 40mb hd, my girlfriend at the time had to teach me to double click the mouse.

Re:Sad this is +5 informative

[matt_hs]

Surely everybody on Slashdot is old enough to remember 386s? What about 286s? I even remember a 186 ... It scared me that where I worked recently with a lot of Gen Y's, some of them may not have ever used a single-tasking OS like MS-DOS or CPM.

I seem to recall the Tandy 1000 used the 80186 processor. This processor came in two flavors, the 80186 and the 80188, same as the 8086/8088. I believe the 80188 had the same modification as the 8088 . . . it had an 8-bit external data bus whereas the 80186 had a 16-bit external data bus.

Re:Sad this is +5 informative

[JWSmythe]

You'll find a lot of the folks here were children, or not even born yet, when those were popular.

    I had my first exposure to computers in about 1980. The school owned it, and had it available to students, but no one knew what to do with it. "Ok, we have it. What now?". So, I sat with magazines that had basic code, typing it in to make things scroll on the screen, and other nonsense. Hey, it was a start, and not bad for a 7 year old who had only seen mini's and mainframes, but wasn't allowed to touch them.

    I think folks should get some exposure to the roots of what they're using now. Is it really unrealistic that someone should know how to open a dos box and get a directory listing, or view the contents of a file without a GUI? Any DOS user could do it back in the day. It's not like you had much of a choice. Now, if presented with a dos box, a lot of people only know how to click the X at the top to get rid of it. Dammit, you must exit your shell, don't just disconnect. :)

Re:Sad this is +5 informative

[ls671]

Damn it I forgot about the 186, I thought it was 8088, 8086, 80286, 386, 486...

On this link, they even state that 8086 came before 8088, go figure !

http://www.pcmech.com/article/a-cpu-history/

Re:Sad this is +5 informative

[hardwarefreak]

Surely everybody on Slashdot is old enough to remember 386s? What about 286s? I even remember a 186 ... It scared me that where I worked recently with a lot of Gen Y's, some of them may not have ever used a single-tasking OS like MS-DOS or CPM.

I seem to recall the Tandy 1000 used the 80186 processor. This processor came in two flavors, the 80186 and the 80188, same as the 8086/8088. I believe the 80188 had the same modification as the 8088 . . . it had an 8-bit external data bus whereas the 80186 had a 16-bit external data bus.

Heh, the CPU in my SMC 7008ABR SOHO router (circa 2003/4) is a "modern" 80C186 (I don't recall the clock speed), running VxWorks.

I too recall the Tandy 1000. In addition to the odball 80186 it had an odball single-sided floppy drive, a 5.25" but storing 720KB instead of the standard 360KB. So, not only did one have a non-standard version of MS-DOS running on the T1000 beast, you had to buy special diskettes which cost 3x the normal 5.25" disks. Did I forget to mention is wasn't compatible with any off the shelf IBM PC/XT software? You could only get software for it from Tandy. You could get all the same software, from Wordperfect to VisiCalc to dBaseIII Plus, but it was all 25-35% more expensive than the IBM counterparts. This one factor led me to buy a Kaypro PC, an IBM XT clone, instead. I was heavily looking at the T1000 because of the hardware specs, but the compatibility issues turned me away.

I later dropped an Intel Inboard 386/PC full length ISA board into the Kaypro, in essence turning it into a 16MHz 80386DX with 1MB DRAM. That was one of the absolutely wildest "CPU upgrades" every devised: A full length ISA board containing a complete 386 system including CPU, memory, memory controller, 387 socket, RTC chip, all the necessary TTL, and a ribbon cable that plugged into the original 8088 socket. The freaky part is that the system still booted the original Phoenix 8088 BIOS and all the option ROMs on the RLL disk controller and 8 bit VGA card (yes you read that right, 8 bit ISA VGA card). Ahh, the good ol days...

Maybe it's his social/PR skills that really count

[shoor]

Various people have suggested that FreeBSD or Hurd would have filled the breach if Linux hadn't come along (and maybe been better). We'll never know for sure unless somebody finds a reliable way to view alternate timelines, but, I'd say there is evidence of non-technical issues in the GNU and FreeBSD camps that could have been showstoppers. Somehow Linux stayed on course. Was that a fluke or something to do with Linus and his personality? I suspect that that was where Linus really made a difference, and without that, the suits might really have prevailed. I think it would have been a disaster if Microsoft had captured the server market, and there was a real danger back in the 90s that that's just what would have happened.

Re:Maybe it's his social/PR skills that really cou

[LOLLinux]

I think it would have been a disaster if Microsoft had captured the server market

Why?

Re: Bill Jolitz and Linus

[TaoPhoenix]

(Father)- Happy Birthday Linus! Go ahead and open your present!
(Linus )- (Opens Blank Email) What the hell?
(Mother)- What did you expect, support?
(Father)- Linus, where are you going?
(Linus) - Out!
(Father)- What are you going to do with your life?
(Linus )- I'm going to code the kernel I want! ...

Code ... Code ... Code ... OMG Code!
Let's write some code.
Let's write some code.
Let's write some code.
Let's write some code!

This code rules.
That code Sux.
This code rule.
That code Sux!

"I think you have too many dependencies."
Shut up!
"I think you have too many dependencies."
Shut up!

Incommunicado Bill... Let's hold a code party!

This code spec has 300 subroutines. Let's write it!

How would the world be different?

[Anonymous Coward]

Well, for one thing, /. would not be filled with a bunch of whiny bitches.

Re:Maybe it's his social/PR skills that really cou

[shoor]

I think it would have been a disaster if Microsoft had captured the server market

Why?

Microsoft's strategy has been to come up with proprietary extras that go beyond open standards. Users get hooked on these and they are proprietary. It locks out competition. If you've seen my sig you probably can figure out that I'm in the anti-Microsoft crowd. (It's actually a paraphrase of Cato the Elder, who was famous for concluding each speech in the Roman Senate with the statement "Furthermore, I maintain that Carthage should be destroyed.") I don't want to turn this into a big long Microsoft rant; there's enough of that on the net already. Suffice to say, I'm pretty damn sure that we're better off not having a monoculture dominated by Microsoft.

Re:get real

[shentino]

As far as I'm concerned he's earned the right to be a douche.

Not that being a douche is a good thing, but IMHO his contributions to open source outweigh that.

His birthday is too close to Christmas like mine.

[antdude]

See http://www.xkcd.com/680/ why... :(

How Linus/Linux Affected PC World

[Velska1]

Linux has undoubtedly had an effect on PC world. I wonder if we can know really what that is, since it's hard to run alternative simulations in the real world, we only have this one history that is all-inclusive, if done right.

Linus is an example of how a project being controlled by a single actor can be very efficient in getting a fairly standardized product out (more so than many commercial operators), meaning, of course just the Kernel here.

In the future, I suppose it should become a committee effort, since there will be too many people, who want to influence it for him to work with eveyone, and too many people getting suspicious of the project being controlled by this Finnish-American guy, who may or may not have political ideas, too... Besides, the Chinese and the Russians already have their own ground-up distros that they use. They probably have government backdoors to every system???

And then the committee members will be bribed and boughtt by megacorporations and they will screw the development in behalf of their employers. Et cetera, et cetera...

Then, one day a guy goes out and buys a very basic computer and build a new kernel for a new kind of OS, because GNU/Linux has become such a f*@$kfest that you'll never be able to predict what's going to happen (if it doesn't end in some megacorp's pocket).

So no, I don't think that the world would have changed a lot. Perhaps Hurd would have gotten more attention, perhaps the BSD legal hassles would have been solved. Perhaps someone else would have created a kernel, and then been able to work with many people to make it better.

BTW, seems the Linus-mocking comments come from AC a lot. M$ fanbois or just people who like to yell "fire" in a full movie theatre?

Hjärtliga gratulationer på födelsed

[TxRv]

I hope that's grammatically correct...

damn cut-off title.

[TxRv]

Hjärtliga gratulationer på födelsedagen Linus Torvalds!

Somewhere in an alternate universe...

[dushkin]

Cartoon flashback ripples...

"Today is the birthday of Linus. Just under 19 years ago, on the first day the shops in Helsinki were open after the holidays, Linus rushed out and spent all his Christmas and birthday money on his first PC: a DX33 80386, with 4 Megs of RAM, no co-processor, and a 40 Megabyte hard disc. Today, the MIND CONTROL SOFTWARE he wrote on that system powers 90% of the world's ROBOTIC OVERLORD SLAVEMASTERS, and is starting to find its way into more HUMAN SLAVE COLLARS — not to mention HYPNO-ROBOTIC IMPLANTS. What would the world look like had he spent his money on something else?"

Pervasive Linux

[DrYak]

and the emergence of internet access at home

Which, by the way is linked to another positive effect of Linux being GPL :
with the arrival of broadband internet, the number of modem/router running an embed Linux has multiplied.

The same has also happened with multimedia -players / -disk enclosure hybrids, NAS, etc.

Linux has got quite pervasive in the embed world (Well, if we're still waiting for the "Year of the Linux Desktop", at least the "Year of the Linux on everything-else-but-the-Desktop" is now past long ago).

Had Linux been under a different license, the situation would be much like the iPhone : there are bits of BSD inside, that's cool, but that's about all you can say and do about it.

Whereas GPL has enforced distribution of the code and better possibility to modify said embed system. Nowadays you find lots of 3rd party open firmware (like OpenWRT), whose quality is ever increasing and some are even starting to get picked up by vendors (some models from Buffalo feature DD-WRT)

The same kind of collaboration as what is nowadays the norm among the "big players" in the server field, is now slowly starting to appear in the embed world.

(And of course, if the companies are attracted to Linux in the first place, that's probably due to its qualities. Qualities that it managed to accumulate, thanks to a license that strong pushes toward collaboration and up-stream contribution, unlike other license which would more favor "each on its own approaches".

Well a benevolent dictator, good at managing the whole project and passively encouraging people to contribute - as you point out - also makes a strong argument for Linux. If this wasn't a requirement, Hurd would have succeeded).

Re:Maybe it's his social/PR skills that really cou

[welsh git]

Mod parent up!

I'm a FreeBSD guy through and through, but it does seem possible that without Linux, Microsoft could have indeed captured the server market, as you say.

Where I worked in the late 90's, they started replacing alot of their working unix systems with more expenesive, less reliable windows servers, where different departments (os support/db support/application support) had to 'book' time on the machines to get work done without clashing with anyone else :-(

Needles to say, they effectively went bust a few years later, but at that time, MS was making real inroads that linux has helped thwart.

Happy birthday, thanks for the present!

[hallux.sinister]

Best wishes for your birthday, Linus! (I am confident he won't ever read this, but I wish it for him anyway.) My computer is one of those running the descendent of what he wrote on that computer. It's almost as if it was his birthday, and he ended up sharing his present with everyone on the planet... kids could stand to learn from this example of selflessness.

The story asked a rhetorical "how different would the world look today..." and the answer is: Gates and Balmer are working on a time-machine, and plan to take one of those hermetically-sealed Macintosh machines back with them, along with the tools required to open it, and give that to Linus for his birthday, and hope he ends up as a hardware hacker instead! ~Hal

Re:over 40

[diego.viola]

Git rocks, I use it too.

Happy birthday Linus.