Linus Torvalds once led a ragtag band of software geeks. Not anymore. Here’s an inside look at how the unusual Linux business model increasingly threatens Microsoft. Also at BusinessWeek, Microsoft Group Vice-President Kevin Johnson predicts the upstart operating system will go the way of Unix”.
Linux Inc. And Fragmentation
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
OK, it can fragment, people can take the GNU GPL liceneced code and make it better the way they think it should be.
But the beauty of the GNU GPL is that the fragments can be recombined as (ie best from each) because they will both still be under exactly the GPL licence, no more and no less.
This is how it differs from Unix and BSD, which can be taken away put under different licences which contradict and can therefore not be recombined.
The GNU GPL is at the heart, let it fragment, compete and recombine… Free-market vompetition, innovation and invention does the economy well.
I agree…nothing like a little vompetition 😉
I don’t necessarily think that fragmentation is a bad thing. Most distros are “hobby” distros and the big boys have held steady as the big boys (for the most part) for quite a while now. Some of the hobby distros give the big guns some new ideas and that’s not a bad thing.
I was thinking kernel forking etc which can recombine… Yeah, I forgot about the ‘distro’ competition too. Because its all GNU GPL then people freely port innovation in one distro to another distro…
PS. by ‘free’ (freely) I mean freedom, not necessarily unpaid!
Lots of companies pay to transfer innovative work to the distro they use at work.
> This is how it differs from Unix and BSD, which can be
> taken away put under different licences which contradict
> and can therefore not be recombined.
which is what happened with Wine and WineX (now Cedega), most suckingly.
> Because its all GNU GPL then people freely port innovation
> in one distro to another distro…
which happened f.e. with RH Kudzu and Knoppix hardware automagic, rockingly…
😉
Lots of companies pay to transfer innovative work to the distro they use at work.
Like the Gnome people paying for code “bounties.”
<<Like the Gnome people paying for code “bounties.”>>
Actually the Gnome bounties are NOT funded by the Gnome Foundation. Novell is paying for them.
such a breath of fresh air over all the hacks doing OS installation reviews and astroturfers spreading FUD
Is it me or at least once a year someone at Microsoft reminds us that they believe that Linux is going the way of Unix and fragment. It could be wishful thinking or maybe they want to make it a self fulfilling prophecy. I’m pretty sure I have heard it more than once from both Gates and Balmer, if not someone else at Microsoft. It very well could be a subconsciousness thing rather than a planned attack. Probably a running theory built on wishful thinking. I agree with that last 3 posters, it will not be a big issue if it did. Heck Red Hat, Suse or whatever have put their own little modified kernels specifically for their release. Especially back in the day when USB wasn’t natively supported in the Kernel. Just a thought.
When I said the last 3 posts I was refering to the first 3. Couple people posted before I got done typing up mine.
The GNU GPL is at the heart, let it fragment, compete and recombine …
How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?
<<How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?>>
Whos post are you reading?!? The author of that comment meant that similar technologies will compete and the best quality ones will survive. He wasent talking about commericalism.
Are you trying to pick a fight?
But since the interfaces to the kernel hardly ever change it doesn’t really matter as far as being able to run a program from RedHat to Suse as long as you have the libraries.
The differences in userland is another issue.
Genetic works well for billion of years by fragmenting DNA code into half parts who can be recombined with other different half parts. The result is that only the best genes of species will be chosen in the long term.
Sexual reproduction was chosen by most of life beings, not cloning.
userland differences? LSB is created to handle just that.
as for evolusion over cloneing, bingo. that is just what the gpl is about. mutations, testing theorys and adapting to new enviroments. and if they give a benefit then that “genetic” code will go on.
I enjoyed this article, I think it made a lot of salient points and was well done.
Linux hasn’t exactly created a revolution in software development but more of an evolution, the internet being the biggest catylist to this process.
Open source development is creating incredible software and while this threatens traditional, closed-source development, it shouldn’t be perceived as a threat. In fact, closed-source, commercial software could just as easily exploit this development methodology to make far superior versions of their own applciations.
Open source development is as natural as laissez faire capitalism itself, an eco-system which when left as little-regulated and controlled as possible (or not at all), can create explosive growth, opportunity, and prosperity for everyone.
Anyhow, excellent piece.
The article is pretty ignorant of the open source movement.
Linus is in charge of the linux kernel, nothing else.
Linus has no inflence over how gcc, openoffice, gnome, kde, or the tens of thousand of small one man projects operate.
I have nothing against Linus, but having the media put him on a pedestal just to make it easy for their puny minds to make analogies to the corprote world does nothing to advance societies understanding of where we came from.
There are still a lot of idealists around who dont code for the money, for corporate influence, or even for end users.
I for one just want to write good code that will stand up to peer review.
Understand it for want it is, dont try and remould it into something you have already defined.
I enjoyed reading it. Most articles these days are written by crappy writers.
I’ll go thru the article from top to bottom and indicate what I think is FUD:
Quote: “We reengineered goals and put in new incentives. We expanded field sales and marketing. It’s all about being connected — listening and responding. ”
Fancy marketing gobbledook if ever I heard it. Basically they’re saying ‘we’re scared shitless of Linux so we’re spending more money on advertising and FUD and donations to SCO to help our cause’.
Quote: “by having the value proposition that we feel is better, we will win customers”
No, by dumming down your operating system so dogs can even use it, more people will buy it. Computers and operating systems are complex, if you dumb them down you WILL introduce security and reliability issues – these factors of ease of use vs security/reliability are inversely reciprocally linked to each other.
Quote: “We can show great examples of how customers can use our technology to run their operations more efficiently”
It’s strange that Microsoft compares bugs/security issues of the ENTIRE subset of Linux kernel & applications that are present in most GNU/Linux distributions, but fails to do so with it’s own operating system, sticking solely to operating system kernel bugs, and ignoring spyware/adaware/malware/viruses/worms that are a result of a dumbed down user experience that has not been done very well. Let’s just compare kernel bugs vs kernel bus Mr Kevin Johnson, and let’s compare kernel bugs response times. Stop being misleading and stop spreading FUD. You only spread FUD because your product is poor enough to warrant doing so in order for it to survive against better products that compete against it.
Quote: “Because of the evolution of the Internet and connectivity, the whole area of security is bigger than any one vendor.”
No shit sherlock. It’s funny though, I remember several key IE exploits that were known for over 12 months and were still not patched!
Quote: “There’s still a lot of work to do, but we’ve made progress. ”
That’s the first thing you’ve got right in the interview Mr Johnson.
Quote: “We have more than 160,000 partners worldwide”
Ah yes I was waiting for this. Let’s equate this to how many independant open src coders we have working on the kernel and applications shall we? I think Microsoft loses out here yet again.
The whole idea of ‘partners’ is companies that are happy to trudge along the Microsoft gravy train because it guarantees money from indirect sales. It is a sole method of Microsoft entrenching itself in anything and everything and ensuring that it’s percental share of any particular market is great enough to become a monopoly and bully any competitor out of the market.
Let’s look at IE as an example – full of bugs, security issues, yet Microsoft has ‘partnerships’ with many 3rd parties who code applications that RELY on IE. Yes, rely on IE. By using these 3rd party applications you are tied to IE and ALL of its security related issues. That’s pretty good partnership!
Quote: “They’ll want to have a unique value proposition, and that will create fragmentation of the different Linux solutions.”
That’s right. Second thing that you’ve gotten right. However, it means BETTER competition for the end user – the customer. Microsoft has no competition, and can innovate (or not innovate) at its own will. Take IE (my favourite Microsoft application to go bashing) – hasn’t pretty much changed in its abilities or offerings to the customer in 5 years now. Big, slower, more bloated and less secure. But no real new innovative features. Looks a bit prettier maybe, but I look more than skin deep.
Quote: “At the end of the day, it has the potential to fragment like Unix did. ”
Ah yes it does! I agree! But – this is because there’s competition and no one given monopoly dictating things. Microsoft doesn’t have any Windows competition, it’s the sole driver of everything and anything Windows.
Unix had a core part (say sys 32, or sys v), and many took that core part and did their own thing, IRIX, AIX, Solaris, HP-UX etc. They all shared basic core parts. This meant better choice for customers, and they used their own proprietary add-on tools to entice customers to their own particular blend of Unix. Linux is the same (draktools vs debconf as an example [i’m ignoring ease of use and prettiness, but concentrating on functionality here]). Windows lacks this because it is a true monopoly. Competition (and competitors going bust) is a part of true non monopoly, such as Unix or Linux were and are.
Quote: ” Customers are going to Windows or Linux. Once all of Unix’ share is wiped out, we’ll continue to gain share.”
So he predicts that once Unix is gone, Linux won’t have any more sales, because the only people switching to Linux are the guys who used Unix? I don’t know what marketing bullshit he reads or uses but that isn’t the case.
This has been on an ‘excellent piece’ as some have put it because it solely shows Microsoft for what it is – a monopoly and a bully. It shows why Unix and Linux were always (and will be) more competitive and innovative. Quality of a product is not driven by prettiness and marketing – it’s delivered by solid good engineering principles and design. For the most part, Microsoft applications lack that good engineering etc.
As an example, Microsoft wasn’t even able to make the Microsoft baseline security analyzer itself, but out sourced it to a 3rd party. Maybe there afraid of screwing it up? For that matter I see very few people using it, very few corporations using it – the vast majority of people don’t even know what it is, what it does, and to be quite frank they don’t give a shit about it [security]. Why doesn’t Microsoft push this utility harder on its website? Because the marketing guys don’t want to encourage the fact that people may perceive Microsoft as not being as secure or reliable as it should be [thus losing sales to competitors who do perform better in this respect]. This is the ease of use factor that Microsoft is engineering into their products – and the direct results it causes are easily seen.
Only the other day we had some poster on osnews.com saying that all real servers should have GUIs! Obviously a Microsoft sysadmin with poor real skills.
Dave
One of the things that I love about Microsoft fans/advocates is that they have an amazing ability to live in denial regarding certain things.
Such as Ballmer, Bill and Kevin Johnson being able to spin any negative or threating issue into a positive and favorable solution, explanation or excuse for Microsoft.
Let me point out the obvious here…
<BEGIN 1ST QUOTE>
“Our mission statement is now to enable people and businesses around the globe to realize their full potential. That’s really about putting customers at the center. We reengineered goals and put in new incentives. We expanded field sales and marketing. It’s all about being connected — listening and responding.
In many ways, customers were looking at how to be more efficient, to drive down the total cost of owning technology. We’re helping customers do more with less and at the same time reduce their application backlog”
<END 1ST QUOTE>
Kevin Johnson, Ballmer and Bill (to name a few) have amazing sales skills and NOTHING more.
Using buzz words and dramatic phrases to communicate only proves that you use these tactics to “support your cause” and that without them your “product” won’t stand on it’s own two feet.
Sales 101!
<BEGIN 2ND QUOTE>
“Look at the facts based on some of the studies we didn’t pay for. They show that we respond to problems more quickly than the Linux suppliers and that we corrected 100% of our vulnerabilities.”
<END 2ND QUOTE>
The worst part about arguing with someone that never learns is that they also never step up to the plate and admit when they are wrong.
Think for yourself and live free or be lead by someone else’s agenda.
“How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?”
Might want to go digging through ESR’s eassys on the subject; basically, the market changes from being product driven (eg, I have a feature you don’t) to more service driven (we created xyz feature, and while others will pick it up, we’re the ones who who are the experts about xyz).
At least my interpretation of the changes to the market foss leads to…
anonuser wrote:
basically, the market changes from being product driven (eg, I have a feature you don’t) to more service driven
Just a note from the economics and marketing department.
Product features are often so-called “search attributes” while services are usually “experience or even credence” attributes. The first can be evaluated before a purchase while the latter can’t. Moving from a product competeting on the former to a product competeting on the later means increased risks for customers.
This brings me to the basic point:
David Pastern wrote:
Quote: “We have more than 160,000 partners worldwide”
Ah yes I was waiting for this. Let’s equate this to how many independant open src coders we have working on the kernel and applications shall we? I think Microsoft loses out here yet again.
Unfortunatly, the number of open src coders don’t help. Industries with products being mainly experience or credence based, tend to rely on “good connections and partnership”: This means costumers often substitute their inability to evaluate the product before a purchase with their ability to evalute the people running the business.
As a result, the argument about partners is valid. If Open Source should be used more widely, we need local businesses that sell Open Source solutions exclusivly. Relying on the brand “Open Source” may even be a competitive advantage right now.
“How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?”
Who told you that competition is good? Software developers will be employed by the government and big corporations, paid hourly wages regardless of code they produce. They will code not for money (a.k.a. greed) but for their own enjoyment (a.k.a. love of coding).
They will be paid, of course, average country wages.
Software will be given for free, as in free of charge. You can share it freely, anyone can post it on the Web, for example.
Only vendor locked-in software will bring profits through services attached. Really free software will result in services outsourced to the countries with the lowest demands for salary.
If Bad Corporation makes you pay for services, per CPU, on annual basis, forces you agree for audits and penalties for non-payment, tells you that in order to get patches from Bad Corporation Network or Bad Corporation Hit-and-Run Software Depository you must pay the hefty fee- it is vendor lock in. It is attempt to poison GPL.
If GPL prevails, the government and corporations will use proceeds from Software Tax to finance software development.
Software Tax: as a part of income taxes, or hidden inside cost of hardware.
This is the right way to profit on GPL.
Does anyone reads GPL Manifesto??? That must be mandatory reading at school, like Marx’s Manifesto was in Soviet Russia. You would not ask silly questions, then.
<<“How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?” & Software will be given for free, as in free of charge. You can share it freely, anyone can post it on the Web, for example.>>
ever try to build a complete Linux OS & desktop from sourcecode (from scratch) that takes a LOT of know-how and patience, and makes getting a professionally built distro sound like a much better idea, (worth a few dineros too)
<<If Bad Corporation makes you pay for services, per CPU, on annual basis, forces you agree for audits and penalties for non-payment, tells you that in order to get patches from Bad Corporation Network or Bad Corporation Hit-and-Run Software Depository you must pay the hefty fee- it is vendor lock in. It is attempt to poison GPL. >>
this wont poison the GPL, Novell, Redhat, and now Mandrake is getting in to the business of offering support for enterprise Linux, when you are running a business with a rack of blade servers and hundreds or thousands of workstations & etc… you will see the value of professional support from a vendor…
but don’t worrry there will still be ISOs for free download so you can run that desktop or small LAN.
there is nothing wrong with business making money and profit from their products & services, this is what makes the world work…
but i say again, Linux will still be FREE & GPLed for personal use so people can have a GNU desktop in their homes…
{insert “In Soviet Russia” joke here)
Fragmentation means nothing in GPL software. As long as its GPL, anyone can be compatable with whatever they want. The only reason UNIX fragmentation was a problem is because it was all closed and proprietary, so one UNIX vendor couldn’t reliably be compatable with another.
In fact, the open nature of OSS makes fragmentation unlikly to even happen in the first place. The current crop of distros could hardly be called fragmented. Almost any linux software will compile and run on almost any distro with zero changes as long as you have the dependancies.
Linux will NEVER end up like proprietary UNIXs did.
“How is GPL software supposed to compete in the marketplace when anybody can simply give it away once they have the source code?”
Well Darius, all that people “giving it away” are competing with each other. About five minutes after the first giveaway the novelty wears off and each one is forced to enter a race to innovate where only the better innovators will be making any profit.
“when you are running a business with a rack of blade servers and hundreds or thousands of workstations & etc… you will see the value of professional support from a vendor… “
When I am running a big enough business, I will ask myself is it cheaper for me, in Russia (or China, or Poland, and so on) to pay $$$ for professional support from some American corporation, or it would be cheaper to hire local expertise.
I can get a good, really good, UNIX/Linux admin for $12,000 year in Russia (much cheaper than that in China, but I am Russian Guy- let me talk about Russia).
Red Hat demands $300-$1,500 annually for server support per CPU: 40 Red Hat servers or more and I would be better off (financially) with my admin than with some guys in North Carolina telling me I should speak English if I need them to help me. What the heck: I am paying them $1,500/year for each CPU and they don’t even bother to learn Russian? What kind of support is it?
Novell Desktop: $50/year? Well, if I have just 250 desktops in my company, honestly- I would rather pay my admin $12,000/year to support them. Is Novell going to explain my accountant (who runs Linux Desktop) how to hook his USB drive to the computer? Can Novell do it in simple Russian terms, during Russian business hours (+7 hours to New York time)?
Well, I see no way for companies to profit on support outside the USA.
Back to the USA: young and hungry Asian economies will soon start to provide thousands of qualified Linux admins for very low fee.
While you won’t, probably, hire Linux admin in Nebraska for $1,000/month any time soon, but for $2,500 (plus H1-B visa)- quite likely. That person will put 12 hours a day working for your company.
So, very soon the only support you will need to sign for will be a support of a box that runs your database. To avoid Linux vendor lock-in, when they tell you: buy one support contract- pay for supporting all Linux boxes with our distro you have,- you can run Oracle on Red Hat Linux and everything else on either SuSe or Mandrake or even free Solaris for x86.
Pay for one server, run hundreds for free (i.e., for the salary of your Linux admin).
Add 100 desktops to your company- no incremental Linux support costs: we know that one Linux admin can support *a lot* of desktops.
“but don’t worrry there will still be ISOs for free download so you can run that desktop or small LAN”
I would rather hire UNIX admin to support 1,000 desktops for which he can download ISOs for free, have him 24/7 if I need him, and support my local economy, than pay $50,000 annually ($1,000 desktops, each $50) for some brand name slapped on top of so-called freedom distro.
Also, if I have my guy, and tomorrow want to switch to another distro- my UNIX/Linux guy can do it for me. No vendor lock-in whatsoever.
I’d like to see how would you approach Novell telling them that because you are paying them $50 per desktop for SuSe Linux they should help you switch to Mandrake because you want to, for no extra charge- after which switch you no longer paying Novell annual fees.
Yes, I see the irony: I am paying for support, one or other way. But my point is: corporations offering Linux solutions will not get money from the support.
So, corporations can either lobby government for Software Tax as it were envisioned in the GNU Manifesto, or try to sell me hardware with the hidden cost of software. Too bad for them, commodity hardware market is already low margin, can’t hide anything there unless you are a monopolist selling mainframes.
Is your admin going to indemnify you, give you a warranty (not against all things of course), guarantee your free distro is going to be there tomorrow (most don’t go anywhere but still…), and all those other nice things?
>>Is your admin going to indemnify you, give you a warranty (not against all things of course), guarantee your free distro is going to be there tomorrow (most don’t go anywhere but still…), and all those other nice things?
I hope that was sarcasm. I have yet to see any closed source companies provide these things. I’ve got piles of CDs from software that is no longer “officially supported”. Some of them are barely a couple years old. (Some for valid reasons–it’s hard to provide such a guarantee when you’re out of business–but I didn’t get a discount or refund, nonetheless.) As for warranties, it’s nice to have a piece of paper to make you feel all warm and cuddly, but if you read them, most software warranties don’t actually cover anything. And if they accidentally do, there is always the clause that the vendor may change the terms of the warranty at any time.
Quite frankly, at least when a Linux distro drops out of sight (which agreed, is pretty rare and usually the result of overbroad commercial expectations: Caldera, Corel, any number of small “I’ll get rich” Y2K stock bubble companies) at least you don’t feel stung by expectations that were nothing more than marketing fluff. More importantly, if a distro does die, it’s rare that there is a problem just using another, similar distro, often with little tweaking. Try doing the same with most closed-source offerings. In many cases, competitors actively program their software to discourage migration to another offering. They’ll let you import from any number of competitors products, but exporting to those competitors is some kind of mystical science. As for IP considerations (which is sometimes used as an excuse) what kind of twit-driven company would allow importing, and balk at exporting to their own format?
I’ll take my chances with a skilled admin and open source over corporate marketing fluff anyday.
I have not RTFA and this is probably OT.
One thing that could help the Linux brand and put a limited but usefull control on fragmentaion would be to create a certifiable “LINUX” standard. LSB is probably right for the task, a simple name change is mabey all that is needed.
The names-“levels” would be comparable according to this table:
UNIX <=> LINUX
*BSD <=> GNU/Linux
Solaris <=> RedHat
Darwin(*) <=> Linux Kernel (renamed from just Linux)
*Could Darwin be UNIX certified?
The biggest problem caused by the “fragmentation” is that the use of the name Linux is so ambigous and many people think that the name referes to a single system as opposed to a system family (as opposed to a kernel, which even fewer people realize).
>Linus has no inflence over how gcc, openoffice, gnome, kde, or the tens of thousand of small one man projects operate.
Actually, Linus does have some influence over gcc, there is cross-pollination between the kernel and gcc when it comes to optimization issues that Linus sometimes complains about
If it affects the kernel performance people sometimes will listen.
God, these idiots will still be calling Linux an “Upstart Operating System” 10-20 years from now.
Sexual reproduction was chosen by most of life beings, not cloning.
Back when I studied theoretical ecology, I read some paper
discussing why there was only two seses, and there was some mathematical proof that this should be better than having e.g three.
So what sexes will evolve in Linux. Could it be free distros like Debian and to some extent Fedoar enterprise distros like RHE and similar stuff from Suse/Novell.