Graeme's

You can get new post notifications through RSS, email, Twitter or Facebook

Is Ubuntu destructive?

Posted by Graeme in Uncategorized at 8:59 am on Wednesday, 29 October 2008

Adam Williamson thinks that Ubuntu has been destructive, unfair competition, and that Mark Shuttleworth could have taken a different approach that would have been less destructive.

Firstly, I question that the competition is unfair. Yes, Mark Shuttleworth is in a position to through money at Ubuntu because of his personal wealth. So what? If he is doing so because he thinks it will make a profit, then that is just capitalism in action.

If he is subsidising Ubuntu even though it does not make economic sense, he is putting money into Linux development and strengthening network effects around Linux — the area where Windows is hardest to compete with.

Ubuntu’s big achievement has been marketing. The combination of Shuttleworth’s high profile and willingness to spend on marketing has got people to use Linux who otherwise would not have done.

Ubuntu has certainly changed the market: it has badly damaged some distros, encouraged Ubuntu (and Debian) derived distros, and hugely damaged KDE by exposing new users to the badly implemented Kubuntu.

I agree with Adam that Mandriva is currently better than Ubuntu: its control centre, easy installation, good KDE implementation and helpful (to a large extent thanks to Adam) forums more than make up for its shortcomings (less dependable package management and lots of bugs in less popular packages). I have used Mandriva for years (when it was Mandrake, starting with version 8), and even paid for some versions. I would pay again if the shortcomings were fixed.

The problem is that Mandriva (and other distros) never got the momentum going that Ubuntu has. They did not do it before Ubuntu appeared, and I do not believe they would have done it if Ubuntu had not appeared.

I do think Mark Shuttleworth genuinely does want to make Linux a real competitor to Microsoft (i.e. with comparable market share), and the sort of aggressive promotion that takes is bound to squash some other Linux companies. Eventually, we end up with bigger and better Linux. Overall, its a win for Linux.

Comments (10)

Comments(10)

Comment by Adam Williamson at 3:43 pm on 29 October 2008 at

Well, you didn’t really address my points.

“If he is doing so because he thinks it will make a profit, then that is just capitalism in action.”

Sure, but it’s still unfair competition.

“If he is subsidising Ubuntu even though it does not make economic sense, he is putting money into Linux development and strengthening network effects around Linux — the area where Windows is hardest to compete with.”

But you don’t answer my ultimate question: what happens in future?

Imagine five years down the road, Ubuntu has killed off Mandriva and all other commercial forms of Linux, and then it tanks because Shuttleworth can’t figure out how to make money off it. It turns into just another community distro. Commercial Linux for end-users is then basically dead.

Are we better off then? I don’t think so.

Comment by Steve George at 7:57 pm on 29 October 2008 at

Graeme,

It’s worth reading more than the Guardian article to really get the full import of what Mark said. It’s one of those where a sentence quote doesn’t do justice to the paragraph comment he gave.

His basic point is that the software industry needs to move from selling the bits (or selling distributing them) to selling services around the software. In a world of (almost) limitless bandwidth this makes sense.

I have sympathy with Adam’s perspective though I would ascribe the value-judgement of “unfair” to reducing the cost of distribution to zero. It’s a typical innovators dilemma where an established business model has been put under stress by a new approach.

I still think that a larger market benefits everyone and that is good for users and companies that provide services and products around Ubuntu.

Steve

Comment by Jef Spaleta at 8:38 pm on 29 October 2008 at

“he is putting money into Linux development”

Is Mark actively funding “linux” development? At any level?

How much Canonical paid developer time has gone into adding functionality to any pre-existing upstream projects? How many projects were started on Canonical paid for time versus as a spare time volunteer project by a Canonical employee?

-jef

Comment by Graeme at 8:40 pm on 29 October 2008 at

Adam, I do not believe that anything is unfair competition if it benefits consumers. That makes answering your other point all the more important, of course.

If Mark Shuttleworth loses interest sometime down the line, there will be a massive opportunity to service the Ubuntu user base. If they are no longer getting it all for free, a good many will be willing to pay. Charge for better repos, sell CDs, sell other products (like Mandriva) etc.

Steve, I agree with your basic point, but I do worry that Ubuntu might become too dominant and shape, rather than just be just another competitor in, the market – see my comments on KDE above.

Comment by Jef Spaleta at 9:21 pm on 29 October 2008 at

Its kind of difficult to service the Ubuntu base… if all of the release engineering infrastructure is in the hands of Canonical and not replicable by the external community.

When launchpad is finally open that will help hedge against a complete cratering of Canonical. But that’s not necessarily the ugliest scenario.

A less drastic scenario where Canonical does find a profitable service model, but one that is not tied to the desktop OS could be far uglier for the Ubuntu community itself. Canonical needs to do more to get more community people into the core development team. So if Canonical has to refocus manhours away from distribution maintenance into a service business profit center.. enough Ubuntu community people will be in place..with the authority to move the distribution forward.

What’s the breakdown between Canonical employees and external community in the Ubuntu core team who deal with the release engineering of the critical packagesets? I just saw that a Dell employee was added to that team. Which is great. But overall when Canonical’s profit center snaps into focus, is the Ubuntu development paradigm including enough community members to replace Canonical manhours if Canonical has to reduce staffing in that area to focus on a profit center?

-jef

Comment by Adam Williamson at 10:02 pm on 29 October 2008 at

I’m sorry, but I just don’t think anyone’s looking at things sufficiently seriously. I see a lot of hand-waving, but no proof.

Steve: you’re doing the same thing as Mark – you’re waving the word “services” around as if that alone is The Answer. What are these mythical services?

Red Hat and Novell have the enterprise ‘Linux services’ market pretty much tied up between them. Good luck to Canonical if it wants to try competing there. It seems like they do, with Ubuntu Server, but I really don’t see that going anywhere. Even if it did, Ubuntu would wind up in the same position as Fedora and OpenSUSE: as primarily a sandbox / testbed for a different product that’s sold to businesses.

If you’re saying there’s some kind of service model based on consumer end-users that will actually work and allow a company the size of Canonical to turn a profit – fine. But you need to explain what that is. And so does Mark. Just waving the word ‘services’ around doesn’t cut it. If it was that easy, they’d be making a profit already.

Graeme: “Adam, I do not believe that anything is unfair competition if it benefits consumers.”

Sure, but the fundamental question is: over what term?

Imagine some company spent a few hundred million dollars over a few years making a *really awesome* distribution. Everyone migrated to Really Awesome Distribution.

Short term result? Benefits consumers. Great.

Then the company said “Welp, this ain’t making us any money, time to split!” and left. Result: not good. There’s no non-company-based structure to maintain Really Awesome Distribution, and it killed off all its competition years ago.

This is not a new idea – it’s the reason anti-monopoly legislation exists. The same thing can happen in any industry; one company swamps the market by providing a really awesome product at a suicidal price for a short time, drives out all the competition, and then screws everyone over. There are reasons why it’s illegal to do that. It may benefit people in the short term, but it screws everyone over in the long run.

“a good many will be willing to pay. Charge for better repos, sell CDs, sell other products (like Mandriva) etc.”

More hand-waving (note the giveaway ‘etc’). People used to be willing to pay for CDs – until ShipIt showed up. Now they’re a lot less so. Not that many people are willing to pay for access to repositories, and besides, setting up a repository network of a quality that people would a) be willing to pay for and b) be happy with once they’d paid for it is harder and more expensive than you probably think (I’ve looked into this before).

Comment by Jef Spaleta at 11:52 pm on 29 October 2008 at

Adam,
I believe the term you are looking for is “predatory pricing.” Trying to make a predatory pricing case concerning an open source distribution stick would make for a fascinating court case. It make for some very informative groklaw discussions concerning antitrust laws.

-jef

Comment by Graeme at 4:49 am on 30 October 2008 at

I am not sure that having an enterprise services business must make the distro a test bed. The same product with and without support would also work (it does work for Red Hat given that most people who buy a license must know Centos exists).

In fact, whatever model they eventually adopt (I am addressing Jef’s point here as well), surely it would be very damaging to destroy the roots of their success.

Also you (Adam) imply that there is no potential for enterprise desktop revenues. It is certainly not here in a big way yet, but it might come in the long term and someone who has his own funding (as opposed to relying on VCs) is in a position to wait it out.

The “etc.” is not hand waving. Take it is short for “all the revenue streams that Ubuntu has killed off”. Certainly, people may not be willing to pay for CDs because Ship It exists. If it went away, they would be more willing to pay. Similarly for other revenues. After all, what are those Ubuntu users going to do if it disappears? Switch to Windows or MacOS, or switch to a non-commercial distro, or switch to a commercial distro and pay?

As for the impossibility of replacing Ubuntu. Everyone else taking part in this discussion knows more about putting together a distro than me, so I am happy to learn, but… Ubuntu is a Debain derived distro and likely to remain so. There are other distros derived from Debian. Ubuntu is entirely FOSS. Some distros have switched between being (directly) Debian based and Ubuntu based (even more than once). Surely creating a replacement for Ubuntu, close enough to keep existing users happy cannot be that difficult?

As for predatory pricing, given that MS does that and worse (without penalty), it will probably take someone who is willing to be that aggressive to take a significant amount of market share from them.

It is very difficult to argue that any software is subject to a predatory price at any level, because of its rather low marginal cost of production. Giving away software free is a long established practice (as is turning a blind eye to piracy) and I cannot recall anyone being pursued by the authorities for it.

Comment by Adam Williamson at 6:27 am on 30 October 2008 at

jef: yes, that’s it – although it’s not quite the exact same scenario, just has a lot in common. Ubuntu’s not trying to kill off all the competition and then jack up prices – it just wants to kill off all the competition and then figure out how to make things work on the fly. I’m not calling for a court case against Canonical, or anything. I don’t think that would make sense at all.

Graeme: “Also you (Adam) imply that there is no potential for enterprise desktop revenues”

Sorry if it read that way – I don’t imply that. There is certainly such a potential – it’s even already happening in certain areas. Red Hat has corporate desktop deployments, Novell has corporate desktop deployments, we have corporate desktop deployments. But what the corporate desktop needs is not what the end-user desktop needs, let alone the enthusiast desktop. Give SLED, RHEL or Mandriva Corporate Desktop a spin: as I wrote on DWW, they’re highly stable and well-supported products…that by enthusiast standards are two years out of date. They simply wouldn’t work right on probably half the brand new consumer systems out there, never mind the fact that they don’t include all sorts of stuff end-users like but corporate purchasing departments either don’t care about or actively do not want.

“After all, what are those Ubuntu users going to do if it disappears? Switch to Windows or MacOS, or switch to a non-commercial distro, or switch to a commercial distro and pay?”

That’s the question, isn’t it? I would suspect a combination of the three, assuming anyone is actually still around and in a position to provide a commercial distribution. I hope we would be, if such a scenario came to pass. I don’t think Ubuntu is going to disappear; even if Canonical somehow went down the tubes, it would become a community project. But that would necessarily work differently from the commercial incarnation. I think it’s most likely that Canonical will figure out some kind of revenue stream that does not necessarily directly involve the consumer edition of Ubuntu as we know it today, and then we’ll see what happens from there.

“Surely creating a replacement for Ubuntu, close enough to keep existing users happy cannot be that difficult?”

Well, up to a point, Lord Copper.

To my mind there’s a distinct difference between community and commercial distributions. To date, it’s mostly been the commercial distributions that take the best shot at putting everything together in a package the largest number of people can use: Red Hat, SUSE, Mandriva, Ubuntu. The major community distributions tend to be either gigantic but slow-moving and lacking in final polish (Debian, Gentoo), or sort of flying by the seat of their pants (PCLOS, for instance).

This isn’t by accident. You can get a lot done with volunteers – the Linux world as we know it wouldn’t exist without them. A lot of work on even the major commercial distros – Mandriva, Ubuntu, SUSE, the lot – is done by volunteers. But fundamentally, by paying people to do this stuff professionally, commercial distros *do* get a better result. It is very hard to get documentation done by volunteers (Gentoo had a good crack at it for a while, but their Wiki is falling into disrepair these days), for instance. It’s not by accident that Ubuntu raised the bar for polish in distributions, when it first came out: it’s because Canonical hired a ton of people to work on it.

Take me as a personal example. I love my job, and that’s a substantial factor in why I keep doing it (it certainly isn’t the pay). But if I weren’t being paid to do it, I simply couldn’t. I work, on average, twelve hour days, split between community relations stuff and technical work. If I wasn’t being paid to do this it would be literally impossible for me to do it, because I’d have to work some other job to live, and I wouldn’t spend the entire rest of my time working on Mandriva for no money (I have to *sleep*, after all). There are a lot of people working for commercial distributors who are in this position.

So, yes, it would be relatively easy – at first – to convert Ubuntu to a community effort, in the theoretical scenario we’re discussing. It wouldn’t suddenly fall to bits (although it’s possible there’d be a big fight over who got to run stuff, and there’d be forks and things). But over time, it wouldn’t be the same project. Bits of it would fall off, or it would grow a gigantic labyrinthine bureaucracy and start releasing every six blue moons or something. That’s my opinion, anyway. Fundamentally, paying people to do stuff gets results; and it’s not at all incompatible with the ideals of free / open source. Remember, free software has never been about everything happening with no money being involved. Or free ponies.

“As for predatory pricing, given that MS does that and worse (without penalty), it will probably take someone who is willing to be that aggressive to take a significant amount of market share from them.”

I watch a lot of anime. One of the big anime cliches in the “boy goes on a journey” genre is the pivotal moment at which our hero realizes he could take advantage of the terrifying dark power within him to defeat the Big Bad Enemy, but it would be a Pyrrhic victory; using the power would turn him into something just as bad as the Big Bad Enemy ever was. (Current example – Naruto). Sorry for the random reference. Anyway. Point is – if you have to resort to Microsoft-like tactics in order to beat Microsoft, is that really a worthwhile victory?

Again, I’m not calling for any kind of legal intervention here. I don’t believe Canonical’s strategy is illegal, either in the letter or in the spirit of the law. I just think it’s not the most productive way to develop the Linux ecosystem, and could do more harm to it than good, in the long run.

Comment by Graeme at 8:47 am on 30 October 2008 at

Microsoft seems to do quite well out of supplying essentially the same desktop to corporate and home markets: as far as I know the different versions just have extra stuff installed or enabled (or not).

Beating MS by using some of their own tactics would be worth it. If one major OS was open source, it would leave the door open to new entrants.

I am not so much thinking of Ubuntu becoming a community distro if Canonical loses interest. The user base should be quite an opportunity for a commercial distro to step in: i.e. do a commercial fork of Ubuntu or make it really easy for Ubuntu users to migrate to an existing commercial distro.

I agree commercial distros produce the best product for most users. That is why I have so far used only Mandriva and Ubuntu: they are the best of what I have tried so far.

Sorry, comments are closed