The Billion Dollar Kernel 289
jesgar writes "The Linux kernel would cost more than one billion EUR (about 1.4 billion USD) to develop in the European Union. This is the estimate made by researchers from the University of Oviedo (PPT), whereby the value annually added to this product was about 100 million EUR between 2005 and 2007 and 225 million EUR in 2008. The estimated 2008 result is comparable to 4% and 12% of Microsoft's and Google's R&D expenses on whole company products. Cost model 'Intermediate COCOMO81' is used according to parametric estimations by David Wheeler. An average annual base salary for a developer of 31,040 EUR was estimated from the EUROSTAT. Previously, similar works had been done by several authors estimating Red Hat, Debian, and Fedora distributions. The cost estimation is not of itself important, but it is an important means to an end: that commons-based innovation must receive a higher level of official recognition that would set it as an alternative to decision-makers. Ideally, legal and regulatory frameworks must allow companies participating on commons-based R&D to generate intangible assets for their contribution to successful projects. Otherwise, expenses must have an equitable tax treatment as a donation to social welfare."
One BILLIOIN DOLLARS (Score:4, Funny)
/pinky to mouth ....
Re:One BILLIOIN DOLLARS (Score:5, Funny)
"Frikkin' kernals with frikkin' lazer beams in their frikkin' code!"
-The truth behind Linux's security
Original Unix License was One BILLION DOLLARS (Score:4, Funny)
Back in the mumblety-80s, standard Bell Labs* Unix licenses came in binary and source versions. Binaries were cheap, source more expensive, universities got discounts so it was nearly free to them. At one point the US Government wanted a license that would give them unlimited rights to the code, because that was what they got for software that they'd paid to have develop, and their contracting bureaucrats insisted strenuously that they wanted that option for Unix as well. The Bell Labs Obnoxious Licensing Lawyers thought about it for a while, decided ok, and gave them a price - One Billion Dollars. The government bureaucrats said "ok, thanks", checked the box on their forms saying it was available, didn't actually order it :-)
* Actually, depending on the year, it might have been Bell Labs, or Western Electric, or various parts of AT the bureaucracy you ordered Unix from changed over the years.
Re: (Score:2)
Honestly, I would suggest, get us a Billion EUR for the European Windows bailout. That would save European markets, citizens and businesses a whole lot of money transferred to the United States.
The EU governments should invest in Linux Desktop, OpenOffice development and so forth to get rid off our Redmond dependency. So far only Munich and Spain gets it while other agencies get corrupted by lobbyists.
They didn't factor in the cost of pizzas (Score:2)
That 1 billion would soon seem like chicken feed!
Taxes (Score:2)
It would be cool if companies involved in open-source development would not have to pay taxes for related activities.
Re: (Score:3, Insightful)
Why should it be limited to successful projects? Since this is open source, even a failed project can be hugely beneficial to society in terms of code, ideas or even just experience. Plus, who would declare success? Would a "successful project" be one that gets 1000 downloads
Re: (Score:2)
Re: (Score:2)
The government should provide operating systems as a public service. That would make a whole lot of sense.
Re: (Score:2)
Re: (Score:2)
Not in charge, just pay the bill, not the Bill.
There are many ways for effective Open Source promotion cmp. Google Summer of Code.
Re: (Score:3, Insightful)
A Linux-powered missile targeting-system? An OpenBSD-based content-filter? A NetBSD-server running identity databases?.. FreeBSD traffic-shaping? Are you sure, you'll approve 0-taxes for all of those — and the "related activities"?
Seriously, as if tax-code is not complicated enough (to the point of harming the economy just by the complexity itself) — exactly by the people like y
I'm not sure COCOMO is a good measure (Score:5, Insightful)
Something based on lines of code like COCOMO is probably not a good estimate for a kernel. Kernel debugging is harder for one. Many of the drivers required some level of reverse engineering as well.
I'd say every "Kernel line of code" is probably worth 10 lines of code in userspace, if not more.
Re:I'm not sure COCOMO is a good measure (Score:5, Funny)
I'd say every "Kernel line of code" is probably worth 10 lines of code in userspace, if not more.
Why? Because you think there's some fundamental difference between low level and high level code?
Papayas don't need to be ripe to be useful. Green papayas can be pickled and be just as tasty as sweet ripe ones. The only differentiation is the time of picking.
Why would you give bonus points to the early pickers just because you don't understand the pickling process?
Re:I'm not sure COCOMO is a good measure (Score:5, Funny)
Papayas don't need to be ripe to be useful. Green papayas can be pickled and be just as tasty as sweet ripe ones. The only differentiation is the time of picking.
What in the fuck are you talking about.
Re:I'm not sure COCOMO is a good measure (Score:5, Funny)
Ah, you're clearly a user-space developer.
I wish I had some shiny pieces of glass to distract you.
Re: (Score:2)
Yes, I think there's some fundamental difference between low and high level code.
Lemme 'splain. Let's say I'm writing sexygrep, which takes the search regex from a computer-connected fleshlight.
In any case, I can code it, try to compile it, fix any compile errors, and try to run. Then I fix any logic bugs/crashes. Repeat until satisfied (or tired out from testing).
If I'm writing kernel code, I'm a lot less casual about it. The edit-compile-test loop is a *lot* longer, for one. But more importantly, there's
Re: (Score:2, Insightful)
Actually, I would say 10 is low. Every bug at the kernel level will be responsible for several orders of magnitude more bugs in userspace. It's not just a question of implementing to spec, it's a question of implementing to spec in a manner that is clear and consistent to every developer using the system.
Re: (Score:2)
Ah, there's the bad analogies we all know and love!
Re: (Score:2)
Brooks, in The Mythical Man-Month, said that compiler coding is about three times as hard as normal application programming, and OS coding is about three times as hard as compiler coding, so your estimate has good precedent.
It could have been the 1000 Dollar Kernel... (Score:3, Funny)
That's the linux kernel alone... (Score:2)
But please, don't use dollars as a metric for that. As soon as any
Ramifications (Score:5, Insightful)
Re:Ramifications (Score:5, Informative)
You'd need to check local laws, but I doubt it: charitable donations are usually only deductable to a registered charity. Mind you, if your local LUG is a registered charity, then you probably could...
Re:Ramifications (Score:4, Interesting)
Re: (Score:3, Interesting)
In the US, you can only write-off money donated to charity, not labor.
So if you give the LUG $500, you can write it off. If you spend a couple days making their website, you can't.
Re: (Score:3, Interesting)
Wait a minute...Am I allowed to write off my FOSS development as a charitable donation on my taxes?
My friend the electrician informs me that when a church gives him a receipt for installing an outlet or whatever, he gets to deduct his labor on his taxes as a gift to the church. Its not such a bad deal for him, if he has nothing better to do at that time, assuming that the church gets the parts donated from a store or the church pays for the parts. Technically I guess he's increasing his liability insurance premium by the value of his gift, and he has to drive his truck to the church, so its not all gra
Re: (Score:2, Redundant)
What your friend is doing is against the law.
Only actual expenses are deductible, never labor.
See here:
http://www.christianitytoday.com/yc/churchlawtaxupdate/judge_donationsoflabor.html [christianitytoday.com]
Which strangely uses the exact example of donating electrical work to a church...
Re: (Score:3, Informative)
No. Donations of time or labor aren't deductible.
Re: (Score:2)
Re: (Score:2)
Because Christians wanted it.
Re: (Score:2)
Because Christians wanted it.
That's rather inflammatory, don't you think?
Would it not also be true that the majority wanted it, and the democratic process put it into action?
Or are Jews, Muslims, Scientologists, etc are completely opposed to tax breaks? No, no, you're right, it was probably those damn meddling Christians!
Re: (Score:3, Informative)
Ostensibly, because charitable contributions benefit everyone and therefore the government should encourage them in the only light-handed way possible, i.e. by not taxing them. Certainly, there are more political answers as to why it has come to be like it is.
You could look at it another way, a charitable contribution is almost necessarily a 'gratis' contribution, as in you receive no quantifiable return for your donation (outside of things like a 'gift' with marginal value). Therefore, it's as if you nev
Re: (Score:2)
Re: (Score:2)
Somewhere there is a form letter for this
You are trying to rationalize the US IRS Taxation System. Your Attempts will fail because...
Re: (Score:2)
Let me take one last serious swing at this. The US Federal tax system is basically one giant charity to begin with. Your federal taxes do NOT (except in very convoluted ways) end up paying for things like police, fire, safety, roads, bridges, etc. Those things are all paid for using State Income Tax (where only some deductions are allowed), and [City|County] [Income|Property|Sales] taxes (where no deductions are allowed.) So the feds say "if you give it to some other charity, the portion of tax you woul
Re: (Score:2)
If you are donating something to charity, why would you want (or be allowed) to pass that cost onto the rest of the taxpayers?
In theory, it is meant to reduce the amount of power the government has. If you don't believe that your tax money is being spent well, you can reduce the amount that you pay and have other organisations benefit from it. Unfortunately, it's open to abuse; people can take away money from the state educational fund and instead donate money to an educational trust that caters for children from their own background. Churches, in particular, benefit. The separation of church and state prevents direct funding
Re: (Score:2)
If you are donating something to charity, why would you want (or be allowed) to pass that cost onto the rest of the taxpayers?
If I give a charity a hammer that I purchased yesterday from the hardware store for $10, the charity has gained a net benefit of $10, and I have lost $10 from my annual income.
The reasoning is that the $10 hammer that I gave to the charity provides a 100% benefit to the charity, whereas the comparable amount of taxes would provide a 15-30% benefit to the taxpayers. In economic terms, as long as the charity can demonstrate it's performing a service that benefits the taxpayers, giving me the choice of wher
Re: (Score:2)
If you are donating something to charity, why would you want (or be allowed) to pass that cost onto the rest of the taxpayers?
Think of it this way:
If you spend an hour doing programming for your favourite charity, that's simply an hour with no income and therefore you don't pay taxes.
If your charity doesn't need programming but does need something else, you can spend that hour working your regular job, but suddenly you can only give the charity maybe 2/3 of what that work is worth -- the rest goes to tax. This makes it more attractive to do things you're fairly bad at for the charity directly, instead of doing what you're best at
Re: (Score:2)
If you are donating something to charity, why would you want (or be allowed) to pass that cost onto the rest of the taxpayers?
Because not all of us subscribe to the theory that all money (or productive output) belongs to the government and that anything we keep is taken from our masters?
A tax deduction is not "passing on" a cost any more than not buying a hamburger is taking money from McDonald's.
Re: (Score:2)
I don't see why not. Charge the open-source project $50 an hour, and then donate your salary to offset the cost of hiring you. Of course, that's a lot of paperwork for no net gain... and it only works if the FOSS is a registered non-profit.
If you're asking "can I work 40 ho
Re: (Score:2)
Am I allowed to write off my FOSS development as a charitable donation on my taxes?
Only if you claim the (fictional) income first. If you then donate the code to a registered charity, you should be able to pull it off. Best case you'll end up not paying taxes for the fictional income.
It's a bit easier to just not claim the fictional income, isn't it?
that is not how capitilism works in the us (Score:2)
I don't get the point. (Score:3, Insightful)
Ideally, legal and regulatory framework must allow companies participating on commons-based R&D to generate intangible assets for their contribution to successful projects. Otherwise, expenses must have an equitable tax treatment as a donation to social welfare.
This doesn't make any sense to me. Since the code has been released as open source, it isn't really an asset of the company that wrote it anymore than it is to anyone else who uses it. It isn't something that could be liquidated to pay off debts, and allowing them to specify it as an asset on their balance sheets seems like just another way to distort the books and confuse investors. I don't see any good coming out of that.
Secondly, I don't see the point in letting them receive tax deductions for their contributions. They made these contributions because it was in their best interest to do so regardless of the tax status. And while it is nice that their contributions help the community as a whole, they themselves are helped by contributions that others have made. If they weren't taxed on the later, why should they get a deduction for the former? Open source is already provides economic and social benefits to those that participate in it's development - government wealth distribution is not needed in a system that already does so inherently.
Finally, even if I did agree with these goals, I don't see how having an estimate of the cost of the kernel as a whole would help - what matters are the specific contributions of the company and there are better ways to figure that.
Re:I don't get the point. (Score:4, Interesting)
The basis for copyright is that the public wishes to increase the amount of work in the public domain. Copyright is a deal between creators and public whereby the public believes that there will be more works generated (and end up public domain) by giving a temporary monopoly to creators. The key in the deal, however, is not to reward the creator, but to generate works for the public domain.*
I would suggest, therefore, that giving (tax) incentives for open source software is in line with this policy. People who contribute to open source are giving up their monopoly rights and their work is available immediately for remixing into new works**. Since time is money I would suggest that anyone who is willing to give up their monopoly period should be rewarded.
This isn't a unique concept: Corporations get all kinds of special and additional tax deductions for various activities such as R&D. We do this with the same line of reasoning: we want more R&D, so we provide an incentive so we can reap the rewards.
Lastly, it should be pointed out that the level of incentive (how rich is this program) should be inversely proportional to the duration of copyright. In other words if copyright lasts longer I've given up more by immediately making it available for remixing and should therefore get a greater incentive. If copyright is short than I haven't given up much and should require less incentive.
* perpetual copyright extension has killed this, but that is another topic.
** Yes, it's not public domain, but they no longer have a monopoly on the distribution of the work.
Re: (Score:3, Interesting)
I don't see the point in letting them receive tax deductions for their contributions.
It's as much or more about showing F/OSS some respect, not just money.
Do you think F/OSS is good? A worthwhile endeavor? A benefit to society? And, so, we could use more? I presume yes to all that. Then how do we get more? Strengthen intellectual property laws even more? Change nothing?
The likes of IBM help develop F/OSS because it is profitable to do so. It helps them sell hardware. But F/OSS is under constant attack from monopolists who fear it as a threat to their way of business and the sys
Re: (Score:2)
No, that's not what is meant by the term. Intangible assets [investopedia.com] refer to non-physical things that realistically increase the value of the company. Copyrights and patents are intangible assets because they can be capitalized, and this ability is exclusive to the holder. If anyone could reproduce an image or re-implement an invention, then they couldn't be claim as assets. If push comes to shove, they could sell their copyrights and patents to pay off debt. Brand recognition is another intangible asset, and again
Tax Credits (Score:2)
What would be lovely is if I could get tax credits for committing to open products that further help mankind in my spare time!
Re: (Score:2)
Lets lobby governments all over the world to make it a tax credit.
Beh (Score:2)
Did they factor in two-hour lunch breaks and the afternoon nap? I guess this calcultion was something to keep amused with as the day goes by.
Salary (Score:3, Interesting)
31,000 euro for a _kernel_ developer?? Probably closer to 3 times that. I know it's an average, but do you really think the maintainer of a memory system, or the scsi stack, etc are worth less than 6 figures?
Re: (Score:2)
On average and probably not full time. Considering kernel hacking is probably (on average) 1/3 of a full job, it's not too bad.
Re: (Score:2)
You usually make the calculations assuming a full time basis. If it's otherwise it'll be reflected in the man-months and that's where the scaling occurs.
As pointed out by others, some places have a much lower cost of living than the USA.
Re: (Score:3, Informative)
If said kernel developers were actually working in Oviedo, the city where they researched this, 31K is more than most would ever make. Your typical graduate in his first local programming job gets 15K at best. 30K is a top level salary over there. Last summer, no local company ever came close to offering me half of what I make in an affordable town in the American midwest.
What a silly estimate (Score:2)
No, what you'd get for a billion Euros is that many lines of code. No idea if the code would be any good. But usually when managers are fixated on the LOC, you get lots of LOC, not necessarily GOOD or FAST code. Just lots of it. Been there, seen it, upchucked, many times.
31,040 EUR??? (Score:3)
An average annual base salary for a developer of 31,040 EUR
What kind of silly number is that? I am 100% sure there is no single person who earns that little... is there?
Definitey not with all the taxes included. That would result in 2299 EUR a month (plus 1.5 months of holiday and christmas bonus.)
Or about 1250 EUR net money on your bank account. Or just below 8 EUR (net) an hour.
As a programmer?? Just... Silly.
That wouldn’t leave you with much, after apartment, food, phone/internet and basic clothing & co. With a bit bad luck (in a big city), you couldn’t even pay for a car. (= expensive fuel)
Re: (Score:3, Insightful)
Depending on the big city, a car isn't a good investment anyway. Quite a lot of the large EU cities have excellent public transport options, respect for cyclists and parking that costs close to that of renting a studio flat ...
Essentially you can pretty much compare most large EU cities to that of Manhattan. You can own a car, but unless you work outside the city it's a waste of money
Re: (Score:2)
As a programmer?? Just... Silly.
No, as a kernel architect, who, to gain parity, are smart enough not to make the mistakes Linus did in his early years. So that you'd wind up with a 2.6.3x kernel at the end, not a 0.9, 2.2, 2.4 or 2.6.0x.
Since the Linux guys are all busy, they'd probably have to go raid Sun for developers - I don't think the world is lousy with experienced unemployed kernel architects. You might be able to divvy up the work among architects and grunt programmers, but at least double the es
Re: (Score:3, Informative)
Even in Romania and even with the whole economy downturn, you can't find a skilled programmer at 8 EUR/h. You can get someone competent enough from 15 EUR/h... and for kernel-level knowledge, much higher that that.
Kernel development is not PHP stuff...
Paving the way for a download tax (Score:3, Insightful)
Though the study only considers the kernel, a starting point has been established. Downloading an entire operating system for free (other than ISP charges) denies the state the revenue from sales/VAT tax that would have been paid on shrink-wrapped product. The downloader receives benefit from the download similar to the benefit received by someone who purchased the shrink-wrap product. Should the downloader be taxed similarly to the tax-paying purchaser?
Now that a value is placed on something that is free, it is ready to be taxed like any other product on the market. What I wonder is, did U of O undertake the study at the behest of the government.
These numbers... (Score:2)
Is that street- or dealer-value ?
American perspective? (Score:2, Interesting)
What you call your 'American perspective', I call brainwashing
Re: (Score:2)
Re: (Score:2)
I wish it were so.
In the USA there is only one entity I'm aware of that teaches in any meaningful sense the ideas of rigorous individuality and non-coercive economics.. that being the Mises Institute.
It is very much the case in the USA that if you want to arrive at the unpopular worldview that pure, unbridled, unregulated capitalism is the only _ethical_ form of socio-political organization ever theorized (and rarely, if ever attempted), you'll have to go looking on your own.
For instance, no Republican actu
Re: (Score:2)
So wait, how come I get modded troll but the OP doesn't?
I demand equality and fairness!
Re:American perspective? (Score:4, Informative)
Re: (Score:3, Funny)
Now who gets the last laugh!
Re:American perspective? (Score:4, Funny)
Now who gets the last laugh!
The other mods who modded you Troll?
Re: (Score:3, Interesting)
Vitriol aside, "Social Welfare" can mean anything, like a organization (say, a Church) in a community providing a non-trivial benefit to said community, while operating as a nonprofit. To put it tactfully, you need your "American Perspective" checked. It improves the welfare of the society (albeit in a somewhat hard to measure way). Saying that society as a whole (outside the open source community) has not benefited from Open Source (to which it pays no material compensation for) is ludicrous, therefore
Re: (Score:2)
I am a little concerned when something done with the power of government can "mean anything". That's a recipie for problems. There are actually a number of people in the US that would argue that Churches and other religiously affiliated entities should stop receiving preferential tax treatment because the benefits are conditional and are distributed "unfairly".
I think this is actually the case with all things done in the name of "social welare". Some individuals receive benefits and others do not, and th
Re: (Score:2, Insightful)
Can someone decode this for me?
Do they want to tax companies that sponsor F/OSS development? Or subsidize them? Or do they want the flexibility to do both, and will change their mind depending on which company and which year we're talking about?
Normally, my in-built translation apparatus resolves "Social Welfare" as "unethical extortion of wealth via the threat of state violence". But that's perhaps just my American perspective..
In the US there are several very deeply entrenched political biases against the responsibility of the individual to society... so yes your background influences how you are taking both the words "social" and "welfare".
Try reading it this way instead;
"Developing commons-based software contribute towards improving the standard of living in a very real way. Most tax entities provide for tax deductions of goods and services to charitable organizations. If FOSS development was given the same tax-reducing benefit
Re: (Score:2)
Ok, so you read it as, the EU should setup tax incentives for companies to release code under F/OSS license, as F/OSS software benefits "society".
It's an interesting point of view for a few reasons. While on its face, I'd tend to agree that "F/OSS is good for the world", it's an interesting thing to actually measure.
One way to naively consider the point is to say that fewer than 5% of computer users are direct benefectors of the linux kernel and all linux / GNU distributions combined, as in, linux has had
Re: (Score:2)
Re:Seems a bit high (Score:4, Insightful)
Re: (Score:2)
Given a dozen developers and a hard spec, Linux could be developed in 8 to 9 months.
I would call that a billion-dollars hard spec then :-)
The value of a program is not in its number of code lines, but in its architecture and in the cleverness of its design. Sure, given a good spec, all you have to do is convert it into line code literally and it may be a short job. But such a spec would be the value of the code and writing it would be an enormous effort.
Re:Seems a bit high (Score:5, Informative)
You are nuts.
12 people at 40 hours a week for 9 months is 1123200 minutes. The kernel is about 12 million lines of code. That works out to a line of code every 5 and a half seconds.
Good luck with that.
Re: (Score:3, Funny)
Re:Seems a bit high (Score:4, Funny)
Maybe they estimated Chuck Norris would do the coding.
Re:Seems a bit high (Score:5, Interesting)
Re: (Score:2)
I'd be very surprised if the Linux kernel code is so badly written that it can be reduced in size by an order of magnitude. It is not exactly your random open source project. Besides, I assume you had the 5000 lines of code at hand when you rewrote the classes? And presumably, you had some kind of framework to test and compare the implementations in. That is an entirely different thing than writing something from scratch.
And by the OP's metric, you'd still only have 55 seconds per line, even if you could hy
you'd be wrong (Score:2)
I suspect that there is a lot of redundant and duplicated code in the kernel. I wouldn't be at all surprised if you could implement it with a cleaner design in closer to 1m lines of code.
You'd be wrong. There's actually relatively little duplicated code in the kernel, mostly due to the fact that it's constantly being refactored. The vast majority of kernel code is drivers and arch-specific stuff.
Re: (Score:2)
Re: (Score:2)
L4 is a minimal kernel, it only covers task switching, inter process communication and context protection. Everything else, including memory management, is done by the Linux kernel, which sits on top of L4 and is just modified to hand off task switching and IPC to L4.
Re: (Score:2)
not yet. But there's always mañana...
Re:And yet... (Score:4, Insightful)
It's all about the apps and drivers - mostly the apps.
It does not matter how fast, secure, reliable, or inexpensive an OS may be; if it doesn't run the apps, it's not of much use.
Re: (Score:3, Insightful)
One hand tied behind your back (Score:5, Insightful)
People who write windows drivers are usually given specs for the hardware.
Given the additional difficulty of reverse engineering, it's a miracle open source drivers work at all.
Re: (Score:2, Informative)
Not all Linux drivers had to be developed from reverse engineering. For hardware typically used in servers like ethernet cards and RAID controllers, the hardware companies often supply drivers and if not drivers at least specs to the kernel devs. Even Broadcom helps with ethernet drivers, it's with wireless that they're difficult.
But with Nouveau, yeah, it's a miracle that it works at all.
Re: (Score:3, Insightful)
Most FOSS stuff is literally just copying things that have already been done.
And that's different from commercial software how?
Seriously, developing software is far more complicated than that. Claiming that X copied Y is a dangerous accusation since it's extremely unlikely that Y came up with something without first seeing the same or similar ideas somewhere else.
Re: (Score:3, Insightful)
I too make my living producing closed source software and yes... coming up with the original spec is work. No, it's not as challenging as working with something someone has already done. Have you been tasked with fixing a bug or adding a feature to code someone else wrote? How about poorly written code with no comments? It's pretty tough. Now imagine you don't even get the code, just an inert chunk of hardware and maybe some pre-compiled binary file. Now make it work
Did you intend to be condescending? (Score:2)
Is open source labor as cost efficient as hiring a real programmer?
Wait, are you saying that Linus Torvalds, Larry Wall, Bram Cohen and Bram Moolenaar are not real programmers?
Or... well, exactly what do you mean by "open source labor"? As I understand it, a copyright license can be open source, as can software* released under an open source license. But I don't know how to extend that to labour---do you mean the labour that goes into producing open source software? If I look at a work process, how do I tell whether it qualifies as "open source" by your definition?
(* an
Re:Did you intend to be condescending? (Score:4, Insightful)
The bread and butter of the open source community are not as high functioning as Linus et. al. A lot of software gets written because it's sexy to write rather than because it's needed. Windows and Mac OS X each have a single window manager and maybe two filesystems; Linux has hundreds of one and dozens of the other because they're sexy and fun to write. We have a half dozen version control systems where MS and Apple each maybe use one or two at most internally. Yet we have few working video drivers. This is a clear benefit of having paid programmers. They write fewer developer tools and spend more time improving existing user-facing stuff, because if they don't, they get fired.
Furthermore, a lot of green programmers start OSS projects to become better at programming. Very little commercial software is written entirely by new programmers. This is why it's hard to stay up-to-date in the Ruby community. A lot of the code is written by new Ruby programmers enamored with language features, and then it has to be thrown away and rewritten differently in the face of real-world demands. There's also more glory in starting projects with promise than in carrying through and maintaining older projects. Few people use FVWM2 even though it's stable, fast and highly configurable. Most Linux users today are probably using Metacity or KWM instead.
Most OSS projects reach a certain level of maturity, get stale and get abandoned, leading to this churn. That doesn't happen in the commercial world because code is perceived as having a dollar value. Sometimes, maybe even frequently, this belief is wrong or overestimated, but it does mean that commercial software is often older than OSS, which (IMO) compensates somewhat for the lack of eyeballs finding bugs. Age finds bugs too.
It's hard for me to imagine the world's most highly performing programmers not contributing to open source, but it's just as silly to expect that they aren't outnumbered by average programmers who don't have time to contribute, or that a dozen average programmers can't produce solid code. In many cases I find they produce simpler, more maintainable code because they're less inclined to the theatrics which are the chief form of compensation for OSS developers.
Re:Did you intend to be condescending? (Score:4, Insightful)
Yea, I wish we could get mother nature to stop that evolution crap. It is a well proven failed model for building quality systems.
Of course he did (Score:5, Insightful)
Come on. This was an artfully crafted troll. Comparing open source to YouTube crap videos, without ever making a direct comparison, yet implying that most open source is like most crap videos: textbook propaganda. Then we have the 'real programmers' line, again implying that open source programmers are not real programmers, without ever stating it directly. Finally, there's the 'twenty experts' line, again, implying that no open source programmers are experts.
Seriously, people pay good money to learn how to write propaganda of that quality. And people who are that good at writing propaganda get paid very, very well. I wonder who 'useful wheat' is working for?
Re: (Score:2, Insightful)
Want to know WHY the Closed source drivers are better?
Closed source driver programmers get the full specs and all details of the hardware including several hardware samples in a test jig setup.
Open source driver programmers get NOTHING. they have to go out and buy the hardware, then buy equipment to reverse engineer it, spend months poking at it trying to figure out how it's supposed to work and then write a driver based on those assumptions.
IT does not have to be that way, it's just that hardware maker
Re: (Score:2)
http://intellinuxgraphics.org/ [intellinuxgraphics.org]
Intel has provided open source drivers and specs for their graphics hardware for several years now.
Re: (Score:2)
There is no legitimate reason for holding back the full hardware interface documentation. NONE.
I always assumed hardware manufacturers were concerned about trade secrets-- like if NVIDIA released all the information on their cards, there might be some information in there that ATI could use to make their cards faster. Is there really absolutely no possibility of that sort of thing?
Re: (Score:2)
Yeah, right. If Linux is so inadequate that IBM and Cray should be interested in using it, then everybody else should shun it too.
Re:So, FLOSS developers; (Score:5, Insightful)
where is your paycheck? Hmmmmm?
My paycheck is in the code. For example, I wrote the Objective-C code generation stuff in clang for the GNU Objective-C runtime. Apple employees wrote most of the parsing logic. I get a full-featured Objective-C 2 compiler that I can use on non-Apple platforms. Apple gets some bugs fixed for free. Both of us get out more than we put in.
Re: (Score:2)
You do know that (some) kernel developers have jobs at places like OSDL, Red Hat, etc where they get paid to contribute to the kernel.