well, they want to say they worth so much more too...
a few years ago their stock does worth that much when IPO...but now...it's penny stock.
that just show you how successful they are replacing microsoft....
well, end up killing solaris and hp unix but not NT server.
How do you count the price tack on backward engineering the unix codes into linux?
Is it just the cost of man hour to steal(backward engineer) the codes or the original development time when the unix codes were written?
I still think the whole linux thing is a backward engineer unix so that it can be free.
there is some original code here and there but mostly borrowed.

Linux is used in millions of devices now as the base OS. I can only imagine the amount of money people could have made off this OS. To bad none of the developers get kick backs.

It's a penny stock because the only way you can make money by giving away software is either A) charge for the hardware or B) sell advertising in that software. Linux does neither and that's why it's a penny stock. B) is the main reason why Google gets away with giving out free software.

There's no doubt that the value of something, were it to be bought at fair market value, would be a particular number. But that doesn't make its value the same as that number.

The value of the labor contributed to America by all of the slaves in the United States has been estimated at $4.4 trillion (source: "Six Meetings Before Lunch," The West Wing), a figure using fair market value wages for laborers at that time adjusted for inflation (in 2002). But the real truth is, were farmers required to pay fair market value for the labor of those men and women, they would not have done so. Something else would've been used.

Similarly, no one would pay $10 billion for Linux. And no one would pay $1 billion for the Linux kernel. You could hire a five-person team of low-level developers to write any kernel you wanted in less than two years for about a million dollars. Totally custom, totally compatible, totally anything. It's not even hard to do, and it's even easier when all you're doing is dealing with hardware and not internal software standards already in place that must be met.

This is a seriously ridiculous report. And I can't even believe it's been published.

Or C) give the product away for free and charge for tech support.

Wait, don't I know a company that does that...

"6.8 million Linux kernel lines are worth about 7500 person-years" averages out to about 900 lines of code per person per year. I would love to be paid $180,000 a year to write only 900 lines of code.

if everything followed the Open Source Model. The Police, Firemen, Doctors, Farmers etc. could all be free and done by volunteers when they have a free moment to work on the problem. The problem is that when YOU need something at that moment under the free system it may not be available. Imagine how much fun it would be to go to an emergency room with a heart attack and finding the Doctor decided to take the day off. However they do have a janitor that once cleaned the heart paddles so he can fill in.
Also I notice that a lot of Hardware companies are using Linux, but not so with Software companies. Why is that? Maybe because it would be a lot harder to steal and produce a piece of hardware then it is to make a minor change to a source file? Under the Open License agreement you have to publish the source code. I program and have been working on a project for a couple of years and getting paid for my efforts. How would you feel if you spent many hours doing work and just before you are able to sell your program “Joe Can’t Program” makes a little change to fix a small problem to your source code and releases it for free? How many people are going to pay for something they can get for free? Look at how much grief the Software/Music/Movie companies get for the high cost of their products.
I just wish that all the supporter of Open Source would do the same in their life’s and tell their bosses “Hey I do not need to get paid because I love just doing this. Besides some one will give me a couple of dollars in tips for all my hard effort”.

I think that the overall trend, and maybe this is being accelerated by advances in information technology, reflects the failure of the overall pay for performance model. How do you justify an economic model that requires everyone to work to eat when we don't need everyone to work? I don't know of a better way, but the incentive model for software and overall technological development needs to change. I wonder if something like the academic model would work better? Teaching is open source by nature.

I think our discomfort with DRM stems from the fact that it makes cultural participation conditional.

Does anyone out there have the magic economic solution that encourages progress while recognizing the value of every individual?

If you have that model, quit holding out. It's time to let your brilliance shine.

"Not everything that can be counted counts, and not everything that counts can be counted."

-Albert Einstein

Not so sure about the teaching model, here in the state teachers unions keep poorly motivated people in important teaching jobs, while continually staging strikes to increase wages that are already ridiculous.

Don't get me wrong, good teachers *should* be well paid, they're educating the country's future. However that assumes the best people are doing the job, and unfortunately unions tend to attract the opposite (why work hard when success is based on seniority instead of performance?).

@yo
You make the same claim that SCO made in court. They lost. You don't know a thing about Linux, so please shut up alright.

@Rick
Your comment is what is ridiculous, not this report. Sure, you can make a custom kernel in a couple years. But "totally compatible, totally anything"? You're nuts. Linux runs on a gazillion processor architectures, from supercomputers to tiny $5 chips, supports a pile of file systems, a huge amount of hardware, protocols, etc. Have you ever looked at the Linux kernel sources? Maybe you should. Maybe it would teach you some respect.

@Hal
The $180000 is the cost of employing a software developer, not the wages he receives. Think taxes, health insurance, etc. And 900 lines of code isn't too bad actually, if they are high quality, well tested lines of code as can be found in the Linux kernel. Developers spend most of their time testing, not writing lines of code.

@Rich
Your comparison of a software product that can be copied for free, and comparing that to services provided on an individual basis makes absolutely no sense. It would be hard to find two things more different in fact. Also, you need to read up on open source licensing. You can write proprietary software running on an open source system just fine. If you do so, nobody can make a small change and give it away for free. It is the other way around: If someone gives something away for free under GPL, no one can make a small change and then close it up. That is what keeps Linux from being gobbled up by a big corporation. This is allowed under the BSD license though, and is what Apple did in making OSX. And by the way, most people involved in the development of Linux are well paid and working for rather big companies.

Using Linux makes very much sense for device manufacturers and everyone using these devices benefits because it makes them cheaper. The manufacturers can focus their attention on the application instead of having to worry about file systems, memory management, TCP/IP, etc. It is a system where everyone gets a lot for giving a little. OS functionality should really be a commodity, and Linux makes it just that. But in itself, Linux has tremendous value as brought out in this study. This is because if it didn't exist, all companies using it either had to invest a lot of developer time to make something similar themselves, or pay licensing fees and royalties to another company. Which you, the customer, would end up paying. In either case, the quality wouldn't be nearly as high as that of Linux, due to the much more limited deployment and lack of peer review. In the end, companies can innovate instead of reinventing the wheel. So I ask, why so much hate?

thank you for bringing a little common sense on this page...

when it will be impossible to hack windows I'll switch to linux. I'm a Windows developer and I didn't have to deal with linux too often, but when I did, it was a nice experience and I intend to just keep in touch with linux cause We may never know when hacking windows will became impossible.

My overall vision about Linux inspires me RESPECT and nothing less.

Yes, I've looked at the Linux kernel source. Even versions as far back as the mid 1990s. And I'm here to tell you that the vast majority of kernel code is of no consequence to the platform it's running on. There are only about a dozen or so hardware abilities which have to be triggered within the kernel to carry out specific jobs. And it is only for platform-specific needs that those portions of the kernel need to be re-written or customized.

I have built a totally custom 32-bit operating system, from the ground-up, not-using-anyone-else's-code or design, not even Linux and its open source. It's was designed around my best understanding of the 80386 architecture. I even wrote it entirely in assembly language (which I would never do again by the way). But, it was a very good exercise and writing it that way actually proved a point: you don't need to have the vast majority of the kernel doing anything special, other than just logic code which operates a particular way based on whatever you decide for your setup. It could be written in pascal because it's just logic code.

Also, two years is more than sufficient to design an operating system which is compatible, configurable, adaptable, and to any platform. And if you stick close to the base hardware's abilities and don't venture off into a lot of software protocols or layers atop the architcture, then it's very easy to have small, core library functions added to the OS which allow future software programs to benefit directly from the work of previous coders - something I don't often see in development today as most everybody has incorporated near-duplicates of someone else's code, rather than re-using the binary in some way afforded it by the hardware.

As to the rest of your argument about Linux running on many devices I say this: If not Linux, then something else. And if not something else, then most likely better speed, performance, power manangement, costs, or whatever other beneficial factor, on the supported devices that do run it.

But apart from all of that, I mean seriously ... how far does one go into the realm of pure speculation? We quickly enter into areas of religion and religious beliefs about what might've been or how it might've happened.

Personally, I really don't even care. But at the same time it's truly an absolutely ridiculous thing to suggest Linux is worth over $10 billion, and the kernel $1.4 billion. It's beyond ludicrous, and it has only accomplished one purpose - people can now cite a finding that gives Linux a number. It can be used in meetings, planning strategies, it can be considered when jumping off the Microsoft bandwagon and onto "greener pastures," etc.

Feel free to contact me if you want to continue any of this discussion. I don't. :) The way to contact me is my first name at this web site. Peace.

Imagine how much tax revenue that would be lost do to the lack of high paying software jobs. The free software model doesn't work unless you have some other ways to make money off it. As others and I have stated.

Don't get me wrong, I use and admin linux.
But's it's not worth 10 billion. It does not matter how much free time you donated to it....free is free. They distribute it for free so no one will buy it for 10 billion. It does not matter you spend your whole life drawing a picture and count every hour as a "worth". If no one is willing to buy your piece of crap drawing...it's WORTHLESS on the market. Linux is a backward engineer unix to start with. Or else how do you explain all the commands are so similar to unix?
"In 1991, Torvalds began to work on a non-commercial replacement for MINIX "
"MINIX is a Unix-like computer operating system based on a microkernel architecture."
" MINIX also inspired the creation of the Linux kernel. Its name derives from the words minimal and Unix.
"
I am just quoting from Wiki.
I guess I am not all that wrong after all.
Just wining a lawsuit only proofs that linux does not copy the code exactly. Backward engineer is not out lawed.

"$180000 is the cost of employing a software developer, not the wages he receives. Think taxes, health insurance, etc"...
yea, right...I think my Fed taxes, health insurance is only up to a 30% of my salary ....so even take off 30%...it's still $1,260,000...still a pretty high afater tax salary you would say?
If any programmer work for me only coded 900 lines of code a year...they will be fired.
I know you like linux...but hey don't get mad.

@Rick
You are right that the vast majority of Linux kernel code is of no consequence to the platform it's running on, since it is there to make it work on all the other platforms. You are also right you can probably develop an OS kernel in two years "if you stick close to the base hardware's abilities and don't venture off into a lot of software protocols or layers atop the architecture". But this last statement conflicts with your statement that "two years is more than sufficient to design an operating system which is compatible, configurable, adaptable, and to any platform". Maybe you meant to say "to any ONE platform".
That shows exactly the power of Linux. It supports almost ALL platforms, not just any ONE platform. This means that as a device manufacturer, I don't have to worry about spending two years to develop an OS to run on my specific hardware, only to find the hardware is obsoleted by the time I'm done. Instead, I can develop my application on any hardware, or on a Linux PC for that matter, and easily port it to any hardware available in the market. The CPU or SOC I'm using is obsoleted? No worries, pick another one, even from a different manufacturer with a different architecture, recompile and I'm done. Another Wifi chip or Ethernet PHY? Load a different kernel module, done. I can focus on my application instead of the platform it runs on. If you made a custom OS, any such change would be a disaster.

@yo
The 10 billion number is for the complete distro, not just the kernel. This includes OS, browsers, email apps, office apps, graphics apps, server software, database engines, everything. If you'd count all the development effort needed to reproduce the 5547 apps in Fedora in a commercial setting, I'm sure you'd easily get to 10 billion. The number is based on how much you'd have to pay developers to reproduce the same work, not on how much someone would pay to buy the whole thing, which can't be bought anyway.
And of course, Linux implements a free UNIX lookalike, but that is a far cry from "stealing" anything as you claim in your first comment. Implementing freely available API's requires no backward engineering. Many POSIX compatible operating systems do that same thing. That's why standards exist. But the "borrowed code" claim is exactly what SCO claimed in court, and they lost. Your claim is unfounded.

"Maybe you meant to say 'to any ONE platform'. That shows exactly the power of Linux. It supports almost ALL platforms, not just any ONE platform."

My point is something you're missing. By designing something that's close to the hardware, meaning there are only those dozen-or-so features that are platform specific and everything else is logic based for operations, means that porting the kernel to any platform means only modifying those few modules which are platform specific.

You are talking about the benefits of Linux as it exists today. I'm talking about the benefits of a properly designed kernel that is not tied directly to hardware, but is tied to common hardware concepts. The bulk of the OS design handles software logic, and the small fringe dozen-or-so modules which are tied exactly to hardware, do their thing to feed the rest of the logic.

It's a very straight-forward thing, but not something that's taught to developer today because there aren't many low-level developers. There are people who aren't considering how to code on a given architecture, but rather are coding for a particular design, which in this case is Linux. And while I admit it is an arguable position that writing code for the Linux is kernel is doing something along those lines, there is a lot of baggage in Linux that doesn't need to be there and it is exactly because of the broad support of hardware Linux supports.

In truth, a properly designed operating system with a dozen or so hooks to actual hardware abilities, can be ported to any new platform (should hardware go obsolete) in very short order, a matter of weeks even. That's a small price to pay for an efficient, tight kernel.

At least, that's my opinion. Peace.

Sorry to pop your bubble but you may want to read the text on GPL. It does allow you to charge what ever you want however look below to see a really big question that has not yet been answered. Just because you believe that someone can’t make a small change to my program does not mean the lawyers agree.

“The GPL in court

A key dispute related to the GPL is whether or not non-GPL software can dynamically link to GPL libraries. The GPL is clear in requiring that all derivative works of code under the GPL must themselves be under the GPL. However, it is not clear whether an executable that dynamically links to a GPL code should be considered a derivative work. The free/open-source software community is split on this issue. “

“Your comparison of a software product that can be copied for free, and comparing that to services provided on an individual basis makes absolutely no sense.”
Let me try to give an example: if you have two people whom work for a living a chef and a book writer, 50 years ago you would have had to pay both of them something to use the results of their labor. Today you can use a magic box press a couple of buttons and get the book writer’s work without having to spend a dime. Just because computers came about should not mean the book writer will not get paid for his effort, but that is exactly what happens when people make copies.

So now let’s get down to how this could affect you. Let’s say me and a couple thousand of my friends come down to where you work and start doing what you do except for free, each of us doing a small amount. How long do you think your boss will keep you around when he finds out he can get the job done free? Don’t think it will happen? Just ask the computer programmers.

So if all the open source people would just follow their ideals they would give away all the things they do free. There is no hate there just a wish people would live their beliefs.

You keep saying that your custom kernel would be coded for a given architecture, and that because Linux doesn't do that, but has broad hardware support, it has a lot of extra baggage. We both agree on that.
But then you insist you can port your custom kernel, coded to a given architecture as you say, to a new platform in a matter of weeks. That's where I disagree. You'd have to rewrite it for your new architecture, or you'd have to put in some extra layers to make it generic enough to support multiple architectures easily. By then you'd be making something like Linux.

Linux sure is heavier then a custom solution would be, that's the price you pay for flexibility. Where I work, we are used to writing our software directly on the hardware, without OS. When we wanted to start supporting networking, USB sticks, etc, we had several options. Either we developed it all in house, which would have been a ridiculous idea. Writing networking protocols yourself is a recipe for disaster. It would take years of testing to get it right. Another option was to buy a proprietary stack that is reasonably portable. This might be a pretty good solution for huge volumes. Some device manufacturers provide free stacks, as long as you use it on their chips. But then we would have been tied to one vendor, which we wanted to avoid.

Going to a Linux based solution made sense for us. With hardware that's just a little beefier, and not expensive at all, we got a networking stack that is as proven as it can possibly be, since much of the internet runs on it. USB support was another useful freebie. Very flexible solution to be able to support USB sticks and Wifi without having to get certified. Then there is the tremendous benefit of being able to develop on a much higher level using Python instead of C. Because all the scaffolding was provided by open source solutions, our productivity went through the roof by being able to focus on our app instead of all the supporting stuff. Linux is doing this for a lot of companies, and is driving a lot of innovation this way. For that reason, it is priceless.

You're not popping anything, I'm well aware of what the GPL says. Applications get linked against a C library, not against the kernel. GLibc, uCLibc, and probably all the other C library options are licensed under the LGPL, which allows linking proprietary applications.

"[...] Don’t think it will happen? Just ask the computer programmers."
I am a programmer myself. Linux allows me to work on making a good product, instead of having to reinvent the wheel. How many construction crews do you know about that feel slided because they have to use scaffolding that someone else made instead of having to rig it themselves? How many would be upset if they could get free scaffolding set up for them? They just want to get on with their building. They need to do a good job with their building to stay competitive.

The only ones who would be in trouble if someone provided free scaffolding would be other people that want to charge for scaffolding. They would have to provide significantly better scaffolding to be able to convince anyone to pay for it. It they can do that, fine, they stay in business. If not, too bad.

That is what Microsoft and other software companies will have to do. If they can make a significantly better product, people will pay for it. If not, they'll go out of business or will need to shift their business. Having open source solutions raises the bar. It forces everyone to do better. Look at how Microsoft made people suffer with IE6 for years, until Firefox came along and stole some market share. Now Microsoft is actively working on IE again. This is a good thing. It is called progress.

I agree that Linux is better because it is here and already does all of these things you're talking about. Still, for new developers the ramp-time for writing simple code in Linux is a few months, and writing significant code in Linux is a few years.

My point is only that the Linux kernel is not worth $1.4 billion, nor Linux anything close to 10 billion. I doubt the kernel is worth $21 million in all honesty, because for that same $21 million, a custom kernel could be developed and ported to as many architectures as there are needs. And while it undoubtedly would not be as pervasive as Linux, its design would merit a pervasiveness because it is conducive to easy porting. And that design could be achieved for $1 million dollars in two years by a handful of developers. And, with an end-added cost of $20 million including hardware once the design is had, it could then be ported in eight months to about 200 different architectures by about 200 developers, each working on custom mods necessary for their needs.

In short, for $21 million and two years you could have the Linux kernel and a ton of functional software.

That's my solid belief. And that's where I stand. If anyone wants to fund an effort for the sake of proving thing thing one way or the other, I volunteer my skills and reputation toward that end. It is, after all, an assessment of what I believe my skills to be having spent so many years writing code so close to hardware.

So you don't mind using an idea ( and not paying for it) that was created by someone who then had it copied and turned loose. If you really believe that then where is your source code? It would be fun to look at and see if in a couple of months there would not be a zero cost copy.

Either you:
1. Highly underestimate the effort involved in what you propose, all the "devils in the details" that will come up, and the perfect foresight that would be required to make porting to different architectures THAT easy
2. Highly overestimate your abilities and those of the 'handful of developers' that would be working on your kernel
3. Truly are a superprogrammer and are able to assemble a team of equally talented superprogrammers. In that case, what in the world are you doing wasting so much talent working at TGDaily? ;)

I really wish you would find someone to fund you so you could do this. Seriously, and I'm not being sarcastic. I would be the biggest fan of the project. The result would be awesome. If it was as capable as Linux, could be ported easily, and would have a lighter footprint, I would drop Linux in a heartbeat. I support Linux because I believe it is the best thing there is for my needs right now. I just want to get my projects done and need scaffolding to build my building. I am very happy that there is high quality, free scaffolding available. But I'm always on the lookout to see if there is anything better.

* Sun bought MySQL AB for 1 billion dollars. MySQL AB is the company that develops the open source MySQL database that runs most of the Web. Maybe you don't agree MySQL is that valuable, but Sun apparently thought so and put down the money. MySQL is only one of Fedora 9’s 5547 application packages that make up this 10 billion dollar estimate. You don't need too many hard-hitters like that to get to 10 billion. Hard-hitters like the Linux kernel, PostgreSQL, the Apache web server, PHP, PERL, Python, GCC, Eclipse, Firefox, Evolution, ... Most packages are obviously worth much less than MySQL, but with more than 5000 of them, the numbers sure add up.

* For all the people whining about all the high paying jobs lost to open source, the taxes missed, yada-yada... The reality of open source is that most of its developers are professionals who get paid very well by their employers to work on these projects. The difference with closed source development is that the people that work on these projects can be working at a bunch of different companies that all benefit from a particular open source project, and are thus willing to invest time in developing it. This prevents a lot of wasted effort with everyone re-inventing the wheel for themselves. The classic idea that open source software is developed by a bunch of unemployed people in their garages is just not true. Maybe some start out that way, but they sure don't end that way if their project is successful.

* The amount of wealth generated on top of open source is HUGE. The whole internet revolution would not have happened without it. The internet fueled open source, and open source fuels the internet. The two are inseparable. Huge server farms for shared web hosting would be prohibitively expensive without open source, and that's how many great websites start. Truly innovative ideas usually start in garages, basements and dorm rooms, not in big corporations. And people working in garages, basements and dorm rooms can't afford expensive proprietary development tools. Give tools to the bright people instead of the rich people, and innovation will explode. What do you think TGDaily is running on? Sure, Linux and Apache, and probably PHP, Python or Perl.

* Arranging bytes and charging for it is an outdated business model. Deal with it. You better bring some real value to the table or become history. Anyone who needs to whine about it is probably still mourning the demise of the steam engine, the horse-drawn buggy, the typewriter, and other industries that once were great in their day.