Fedora 9 Would Cost $10.8B To Build From Scratch
ruphus13 writes "The Linux Foundation's recently released report claims, '... it would take approximately $10.8 billion to build the Linux community distribution Fedora 9 in today's dollars with today's software development costs.' The article states why this might actually understate the value of the distros, though, since it doesn't include the power of the brand and the goodwill value. 'There were several approaches that the Linux Foundation employed to reach the $10.8 billion dollar figure, including calculating the number of lines of code in Fedora 9 (204,500,946), and using an average programmer's salary of $75,662.08 — as determined by the US Department of Labor — to measure development costs ... On the balance sheets of Coca Cola and many other huge corporations, you find goodwill listed as a major asset.'"
Holy hell, that's probably more than I make in 10 years.
They can spend twice that much money and only deliver a tenth of the functionality! That's a win! I think.
Kwisatz Haderach
Sell the spice to CHOAM
This Mahdi took Shaddam's Throne
Goodwill only shows up on the balance sheet when an acquisition or some similar event occurs which creates a discrepancy between the purchase amount and the balance sheet of the acquisition.
You don't just make up a number and add it onto your balance sheet.
...is like paying airplane manufacturers by weight.
[...]to build the Linux community distribution Fedora 9 in today's dollars[...]
I'd rather build it in C with a modest compiler.
Otherwise whatever corporation would have created it for 10.8B would have needed to sell it to nearly 200 million people at 50 bucks a pop to have broken even. Most likely they would have folded.
What a ridiculous sum pulled completely out of thin air.
...a similar estimate for the kernel alone. Or, perhaps a more generalized number which would take into account other distros.
If your only tool is a hammer, every problem becomes a nail.
I wonder if I can spend some of my karma down at Taco Bell for a burrito?
A work that expires before its copyright never enters the public domain and thus enjoys eternal copyright protection.
According to an Inquirer article, the estimates were about $10 billion.
I would think it would cost more than $10.8 billion to develop FC9 from scratch then...since it's a better OS.
I bet if you put the specs on eLance, there'd be a company in Romania somewhere bidding to do it for about $427.33, give or take a few dollars :)
yeah we should put a cap on that price
I don't mean to sound cynical, but this calculation seems about as contrived as the RIAA's "billions lost to piracy" numbers. $11 billion?
Also, if that's all it cost, why hasn't Microsoft made Linux yet?
(-1, Raw and Uncut is the only way to read)
Man, I'm in the wrong business... :0
When the going gets weird, the weird turn pro. ~~ Hunter S. Thompson
Kind of glad I saved my $10.8b and got my OS off bittorrent.
You make $3.78/hr ?
Or twice that per hour and half the hours, while working part-time to grind experience. Worse, I had to work for two years for $0/hr as a regular volunteer for the local veterans' hospital before employers would see past the symptoms of my (mostly controlled) mental illness.
Thats not a fair comparisons of cost.
Especially since you are comparing lines of code in OSS to lines of code in CSS, Its like comparing 2 fruits, they are close, but not the same.
---- Booth was a patriot ----
Fedora 9: $10.8B
Linux Ecosystem: $25B
Free Software: Priceless.
Goodwill only shows up on the balance sheet when an acquisition or some similar event occurs which creates a discrepancy between the purchase amount and the balance sheet of the acquisition.
If "goodwill" is not the right word for the value of intangible assets such as trademarks, what is?
Ten billion for an operating system... am I the only one thinking that the money we spend on military adventures and bailing out Wall Street would be better spent funding the creation and development of open source software?
It sounds like a lot of money at first, but ten billion dollars divided by infinity is still zero.
Fedora contains a lot of redundancy. throwing in several text editors makes sense if they're already there and free, but you wouldn't rewrite emacs, joe, Vim and nano. You wouldn't rewrite Epiphany, if you'd rewritten Firefox.
The number's a lot bigger than it needs to be.
One more pun like that and they'll never find where you were beret'd.
which is totally what she said
I was reading the rss feed and thought it said $10.88. I thought "Wow that's really cheap!". 8|
> On the balance sheets of Coca Cola and many other huge corporations, you find goodwill
> listed as a major asset.
"Good will" is a specialized accounting term when used on balance sheets.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
So the average programmer makes 75K/y and it is 10,800,000K that means that they estimage that it would take approx 144,000 programmers 1 year to code 204,500,946 lines of code or 1 programmer 144,000 years to code 204,500,946 lines of code... ... Which means that each programmer delivers and grand total of approximate 1400 lines of code/y. Not very impressive. I would think most programmers could do that in a couple of days.
To me it sounds like someone forgot on or two zeroes somewhere.
Dude, fix your bots or turn them the fuck off.
If I want to pay for software I'll buy it , I don't want to have to start paying for it through friggin taxes FFS!
i bet it wouldn't be anywhere near bowler, colorado
I heard Gentoo linux has some sort of autobuild tool that builds everything for the system from scratch, I think it's called e merde or something like that(you can google it yourself if your interested).
Soundproofing Acoustics noise
I'm not really sure what they are trying to show with this. I'm sure that MS could go and roll up all of there cumulative costs for XP, Vista, or Windows 7 or whatever and show a worthless number that is much more impressive. And they are able to do it as real costs to them and still make a profit, while employing thousands of people. I'm not real sure what this even includes. Are they counting the source to every package that is built into Fedora, ie emacs, all the gnu utils, etc. Are they trying to point out how people are idiots for contributing for free something that Redhat will now tout for their own purpose and profits? Should they advertise it as, people all over the world have saved us billions of dollars to help us profit? What's with the recent posts with really big useless numbers later? Is the Linux community feeling inadequate in a certain area?
When reading this, I couldn't help but wish _I_ got paid that much money. The figure I get from sloccount ($55K) seems high, but this is even higher than that! Anybody want to make me an offer?
Please correct me if I got my facts wrong.
Duh, they are talking about the value of the source code, try getting your hands on that!
-- Lattyware (www.lattyware.co.uk)
In many ways, this exorbitant expense represents a terrible failure of the way we do computing. If it would require such an enormous investment just to get our simple digital machines to do something useful, then there is a major problem somewhere.
Computing has gotten far too bloated and wasteful. The typical "desktop" environment is totally unnecessary and its reduction or elimination would allow a much simpler -- and cheaper -- approach. My Linux system is highly minimalistic. (Thankfully Linux does provide that choice.) I don't use either Gnome or KDE, and I shun most GUI front ends, but I can still operate quickly and efficiently for everything that I do.
The basics are cheap. It's the needless accouterments that drive up the overall cost.
just as useful as the number microsoft gives me for how much TCO is for linux.
give me an independent third party.
Good people go to bed earlier.
Now, now. There's beanie 'nuff of that talk for one day.
...and you still suffer dependency hell. Drop RPM's for deb's, and I'll use Fedora.
just because something is cost X to make doesn't mean its worth Y. Yes goodwill is an asset but coca cola's goodwill represents hoards of paying customers while linux has not.... For 10 billion dollars I'd expect a few things like how about auto installers?, ive had linux geeks say its impossible and its a windows thing and linux will never be windows, but its more like a modern commercial OS thing in reality - I could have sworn my amigaOS in 1987 had auto installers ( If you had a hard disk...fuck that was a long time ago now... ) and so does mac, so linux geeks, how is it just a windows thing? say its currently impossible - fine, dont say its a windows thing... This also doesnt mean im not pro-linux... i am def pro linux, but i cant ditch windows and use it because I am a pro musician / gamer and linux cant support my audio HW (dont say wine, im a musician not an I.T genius), or give me ports of the best software or give me the latest games...I want to see real money thrown at linux, I just cant see it meeting my computing needs and competing with windows on a real scale until a company turns it into a real product that they sell and make money from. how much does it cost to bribe HW makers to start making drivers and software developers to start porting the best software these days? being 100% free is indeed awesome but its a double edged sword.
Goodwill is an accounting term used to reflect the portion of the book value of a business entity not directly attributable to its assets and liabilities; it normally arises only in case of an acquisition. It reflects the ability of the entity to make a higher profit than would be derived from selling the tangible assets. Goodwill is also known as an intangible asset. [...] The difference between the purchase price and the sum of the fair value of the net assets is by definition the value of the "goodwill" of the purchased company.
(Quoting Wikipedia so of course the true meaning might be something like "a giant catfish from Pluto named Horace, who likes to eat acorns", although this tallies well with other sources.) Anyway, it can imply that the company's reputation etc. increases its value above its "paper" value, but it generally means any nonspecific, hard-to-transfer value in the company which is not directly quantifiable (i.e. not intellectual property).
No kidding!!! What do you say at this point?
God, all that good money to build such a shit !
Yeah, you're right. I wouldn't want to be arrested for dis-turban the peace.
which is totally what she said
Going by this metric... how much did it cost to develop windows? How much did it actually cost?
You mitre-gret it when you found yourself in such serious trouble. You should have a chapeau-rone!
What do you mean by auto-installer?
Fedora (what this article tries to value) has had an automatic, unattended OS installation mode for many many years. It actually was in Red Hat before and pre-dates Fedora. Google "fedora kickstart" if you want to know about it.
There is also the yum/RPM system if you are talking about installing applications. This is very similar to the "install wizards" that are used to ship Windows software. What you have to understand is that application software vendors put effort into packaging up Windows apps to have that sort of install wrapper. The OS doesn't somehow do it for them. The same is true of Fedora apps (or Debian). Someone has to do the labor to package up the app binaries to make them suitable for install.
So the right question was how much will you pay someone to repackage apps, if your complaint is that some application you want is not already available in Fedora. I've found that Fedora plus one of the alternate repositories such as Livna or ATrpms gives me all the OSS apps I would want, so I never ever have to download sourcecode and do any manual builds or installs. The only compiling I do is for my work as a software engineer, not as a user of my laptop or workstation. For me, software doesn't exist unless it is already in the repo I am using. It isn't worth my time to download anything else and fiddle about.
I'm pro-linux and all, but no software is developed in a vacuum. Even commercial code is shared between different software and bought from different sources, bringing down the cost of software development considerably. If you want to compare numbers, you have to take that into consideration.
you can't propel yourself forward by giving yourself a pat on the back.
Without reading TFA, I can say that this figure is completely ridiculous and was obviously pulled out of someone's ass.
Let's assume that a top-notch software developer costs a company $250000 a year (including taxes, location, equipment etc). This is actually rather on the high side.
At that rate, you could hire 1000 developers for 40 years. Not to put Fedora or any other Linux distro down, but it does *not* take 1000 developers 40 years to re-implement *everything* in it. No way.
then don't build it from scratch...duh
FYI there is a typo on the screenshots page- under the 'Childsplay' screenshot it states "GCompris is another suite of educational games"
I really like the idea and design, although my daughter (no Asperger's) is still a few years from using a computer. Keep up the good work, and I'll revisit the site in a bit!
Through the years car manufacturers have put "Handling tuned by Lotus" on their cars as a selling point. Lotus has immense goodwill on the handling front, as the name is practically synonymous with excellent handling.
Thus the goodwill for Lotus is that manufacturers will continue to come to them for handling expertise (much of Lotus' business is engineering contracts). The reputation (goodwill) ensures repeated business for Lotus and increased sales for their customers who advertise the Lotus involvement. Less money would roll in without that reputation. The company is definitely worth more with that reputation than without.
So the reputation/goodwill must be worth something. How much I leave to the bean counters.
The comment at the end of the post, "On the balance sheets of Coca Cola and many other huge corporations, you find goodwill listed as a major asset." irked me. In accounting terms, goodwill isn't the warm fuzzy feeling that corporations hold for us. Goodwill is simply the difference in the book value of an acquired company and the actual sale price. Since the assets of the acquired company need to be added to the balance sheets of the acquiring company at book value, the difference between book and market price needs to be held as a seperate item to keep the book balanced.
The comment at the end made no sense whatsoever and goodwill doesn't even seem to be in TFA, so can anyone explain why it's in the post?
You go to wikipedia for the definition, and then criticize wikipedia for not being reliable. If you personally consider wikipedia to be a good source for the definition of goodwill, then don't denounce it, appreciate it.
That is exactly what happend with proprietary software, the selling price of competitive software is often too low to make sense writting it. Now, when you don't writte the software, you never know what is its potentital, even if it could end up worthing bilions after a few iterations. Competitors are also priced out of the market, because their earlier iterations aren't any good they don't have money to take over the world.
This way, the market becomes too dumb to fullfill most of the ninches, fighting only for the top seller softwares.
Rethinking email
I know we're talking about completely debugged code, and including all the design/architecting costs, but that's a pretty high cost... and at $75k salary, it only equates to 6.5 lines of code per working day, or 1450 lines per year!
Hell - I write more code that that on average over the year, and I'm in MARKETING. :)
MadCow.
I used to have a sig, but I set it free and it never came back.
Dependency hell is not a feature of the packaging format. To make it go away, you'd need a huge team of developers focused only on making your distro work, and the proper environment.
Even Debian testing sometimes has dependency problems.
Rethinking email
Anyone else read the title as $10.88? I saw it pop up in my RSS and couldn't believe it. I guess it pays to RTFS
First of all, take into consideration ALL the lines of code that came from Unix as a given freebie
that should not be calculated, remove all the vendor drivers created by the vendor and not the OS people, as this is not included in the cost of knowing how Windows costs, so why do it in Linux as well (unless they have rewritten them...)
Also remove all the sofwtares that come into play that come from third party but are not really part of the OS. Let's say KDE or Gnome or Grub or Lilo. These were made for any distro by xyz person not currently employed by Fedora to actually cut a paycheck for...
Now how much we talking about.
Don't get me wrong, I am a penguin lover, but I just hate to see that now Fedora is getting in on the blinders game of misrepresentation of their products that we have gotten used to from M$
We don't need another M$ machine, stick with being good guys, loving their penguins(whatever flavor)
That's proven. I'm sure the bean counters have a way of figuring out how much. I'm not a bean counter, and I'm guessing you aren't either, so we really aren't qualified to determine how goodwill would be valued.
Most of the development can be outsourced, that would cut the "average salary" to at least 1/3.
--
Stay tuned for some shock and awe coming right up after this messages!
Totally ridiculous! Insane and out of this world!!!
How a software project can cost more than 10 BILLIONS?!?!?!?!
Pure thieves, and the company that pays 10 billions to develop a piece of software must be composed by abnormal and stupids.
Give me a break for God's sake!
So, if you minus the lines of code that RedHat/Fedora didn't write (upstream projects developers' code), and added in the amount of man-hours associated with packaging/debugging/patching these packages, I wonder what they'd be at.
USD 3.50 or so :-)
Joking aside, they did contribute quite a lot in the early day, including inflicting RPM on the population :-)
Insert
am glade moderators has great sense of humor, joy
"Steve Jobs invented the world" -- Bill W. GATES
That's an interesting calculation, but it ignores the fact, that money can't buy everything.
Sure, you can hire a lot of devs and start writing the system, but the can't get it right the first time.
FC evolved over a long period of time and a lot of testing and learning from past errors where involved until they get where they are.
Some parts are also implemented by specially talented uniqe people hwo just aren't there anymore, and you can't hire them too.
Also FC (or any other peace of software around) is not just the code, but also the history. You can't buy history.
Likewise you can't buy the userbase and the developerbase that use / contribute the system. You have to invest a lot of effort and money to do this.
Now if you say, that you want to create an OS with similar functionality and stability then FC and start to calculate, go on, but the figures will be quite different.
Does this mean I can pay my mountainous student debts with good will?
pretty please?
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
How much money would it take do make a _decent_ disto...
More code does not equal actual worth.
While it's true from the standpoint of "if we were to write this now how much would it cost to pay developers, etc.", which I know was the point of the report, that just doesn't equate to anything substantial on its own.
Really great things don't necessarily require a lot of code (think JUnit and other similar libraries), and there are a lot of projects and products that had a *ton* of money spent on development that were flops or not so great. While it's true that an OS is obviously a lot more complex than a Java unit testing framework, it doesn't mean that an OS that has 1/2 as many lines of code would be of any less worth than Fedora. I'm not knocking Fedora, but I just think that this whole report is just going down the absolute wrong road, and it scares me to think that anyone in the Linux community would even think it worthwhile to post this kind of stuff. It just makes me think that people have lost their way from the "path of enlightenment".
You are correct, but incomplete. Goodwill is generally amortized - that is, a portion of it is deducted against revenues over time. The rationale is, you paid more than book value for the assets, and over time the additional payment 'causes' revenue to rise (otherwise, why would have paid more?) so you offset those revenues with the goodwill. It's a makeshift way of accounting for a business judgement (i.e., to pay more than something is 'worth' in a strictly-book sense). It is relevant because such amortization reduces operating income, which reduces taxes. Recall, corporations pay taxes on what they net - not (like US Human persons) on what they gross.
Goodwill is not thie TFA, but is in the post because (I think) there was discussion about the value of the brand in the Fedora product under discussion. Goodwill is only an issue if someone actually bought Fedora from RedHat - you don't get goodwill without a real transaction, and you don't get to 'make up' the number either - it is the numerical difference between what you pay and what is already recorded on the books of the seller. The only impact you can have on goodwill is in terms of what you are willing to pay for it. If you are wrong, then you end up in a situation with 'impaired goowill', which requires writeoffs, and can seriously impact your net income.
Seriously, how much of the code of fedora 9 is redundant?. Your average distro is a big collection of gpl code from various smaller projects, and I assume the 10 billion figure is the calculated value of the entire code base. Cutting down on outright duplicated effort and cutting out things that can be left to 3rd parties to develop for your product you could code from scratch an equivalent OS for MUCH less. The only part that would be somewhat unshrinkable would be the kernel, but even that i be simplified.
So cutting off all the fat, I think you could build something close to fedora 9 for under $1bn at full price. Just look at how much functionality can be packed into some of the mini distributions going around, when you choose just to have one browser, one mail client etc etc. That's a clue.
After logging in slashdot still does not take you back to the page you were on. It's been that way for 20 years.
The average line of code in industry is of much lower quality than the average line of code in FC9. Some of the GNU utilities and libraries have been polished up for 20 years, the kernel for 15, X for 20 years, etc. Code that is on its first release is often in production use in industry, but seldom used in any of the big distros -- that's the *other* 100 million lines of code sitting out there in the free software world. So I guess that $10.8B is low by a factor of two or three compared to the money it would cost to develop such a polished product if Microsoft, IBM, Sun, or Oracle were to attempt it.
"If you're not passionate about your operating system, you're married to the wrong one."
...really include p0rn-comfort?
There's no failure quite as dissatisfying as a complete and total solution to the wrong problem.
this is a complete load of junk. this must be some backhanded effort to try and boost Fedora's or Redhat's stock up.
anyone writing this completely misses the point of a distro and how a distro is installed on a PC.
Fedora/Redhat didn't write the kernel, they didn't write any of the GNU apps, they didn't write X, they didn't write openoffice, they didn't write any of the 3rd party apps that make the distro complete. And nor would anyone.
that's the problem with pulling stuff out of your arse -- you end up with a handful of shit.
The reason girls and Windows users don't understand UNIX is because all the documentation is in Man files.
If I do the math, $10.8 billion dollars divided by 204,500,946 lines of code I get a little under $53 per line of code.
Studies of software development productive frequently come up with a number like 10,000 lines of code per programmer man-year. A recent large project I read about at Cisco, doing system level code, I seem to remember saying more in the 7500 lines of code range per programmer man-year.
If we do the math again, 204,500,946 lines of code divided by 7500 lines per programmer man-year, gives a little over 27,000 man years. If we take the claimed cost, $10.8 billion and divide by 27,000 man years, that gives $400,000 per programmer man-year.
Also for a little reality check, Microsoft's R&D budget is around $8 billion/year now, so Microsoft would have the financial resources to write the equivalent of Fedora 9 every 16 months.
Could this possibly be why we don't see any new major Linux distros built from scratch?
Why don't they post something about how much Windows 7 would cost if MS developed it from scratch?
BS. goodwill can only be listed if it is bought - ie buying a firm with $1m physical assets for $2m - $1m is goodwill - WHICH STILL DEPRECIATES. Of course maybe US accounting standards allow companies to make up a figure whenever they want. In which case my company, which I just invented 3min ago, has a goodwill assets worth $10m and therefore has total assets worth $10m and $2 (I donated my pen to it).
So, by the same calc how much is Windows worth?
What a ridiculous sum pulled completely out of thin air.
It's not out of the question. Given current software development cost models, the Linux kernel will cost about a billion US$ to redevelop soon http://www.dwheeler.com/essays/linux-kernel-cost.html
Given the fraction of the code base of a complete system that the kernel is, I'd say US$10.8B if anything is on the low side, if one were to do it again from scratch and pay for the development.
My crap worth $30 billion, but who's gonna buy :(
Thanks for spotting the typo, it's been corrected. That's what happens when you try to put something together from scratch in a day.
How old is your daughter? I have a 3 year old who absolutely loves the computer I setup for her and her 4 year old brother.
http://www.mhall119.com