Ubuntu's Shuttleworth: "I don't think anyone can make money from the Linux desktop." | Hacker News Reader

Ubuntu's Shuttleworth: "I don't think anyone can make money from the Linux desktop."

blogs.computerworld.com

29 pointsqhoxie35 comments

Mozilla is making millions on advertising from Firefox search and I think the same business model can be applied to Ubuntu. For example, by adding a search bar directly on the desktop.

That's still not making any money on the free software itself or from the users; that's just another "added cost service", and the persons paying for it, instead of being the user, is the one wanting to advertise to the user.

I could just as easily use any search provider in Firefox that I want (and I regularly use search plugins for Wikipedia, IMDB, Python/PHP references, Thottbot, etc) in place of Google, and then Mozilla no longer makes money from me.

However, that's not to say that your idea isn't an excellent one. I'm just pointing out that even if that idea had been implemented, Mark is still right. ;)

Better yet, since they are free to make modifications to Firefox, they could always change the account to an Ubuntu one rather than leaving the Mozilla one in there for search.

Wouldn't be so nice to Mozilla, but that is the nature of open-source. In fact, if you have a company with lots of employees who use Firefox, it might make sense for you to alter that. Of course, probably isn't worth the effort considering how little you'd make back.

They're not free to make that sort of modification to Firefox. Patches for Firefox (and a number of other Mozilla products) need to be vetted by MoCo, and I doubt they'd be approved.

They are, of course, allowed to rebrand and then make all the modifications they wish.

This was the impetus for the creation of Iceweasel.

Take Google away and how much money does Mozilla still make from the search bar?

Deep question: Take away Microsoft and how much money is Google willing to give Mozilla for the search bar?

Take away Microsoft and Yahoo and the answer is $0.

Good question. But then, Google is about as dominant as we can expect anyone to be long term.

For the foreseeable future search is worth a lot. So for the foreseeable future, helping search increase market share (even if it's only by 0.X%) is worth a good bit too.

Actually, from a simplistic viewpoint, the ff search deal is worth more to MS then it is to Google. If FF didin't strike that deal, what engine would they put there? The cost of putting up MS/Yahoo in terms of user experience & pundit complaints would be something compared to Google.

Also, Google would probably end up with a lot of that search traffic anyway. MS wouldn't have. Google is in a snese paying for traffic they may have gotten anyway. MS would be paying for nee traffic.

"we have no plans to create a traditional desktop product for the consumer market in the foreseeable future." -- redhat, 2008/04/16

http://www.press.redhat.com/2008/04/16/whats-going-on-with-r...

The question is not who makes money _from_ linux on the desktop, the questions is, who makes money _because_of_ linux on the desktop.

Canonical's profits are going to come from training, certification and consulting partnerships; but they aren't the only ones who will be making money from the existence of capable linux desktops.

Any org that has a large number of client installations would probably save money (easiest way to make it) by implementing a linux desktop.

Is there any big player still trying to make money off a Linux desktop? I thought this issue was resolved long ago. I agree with Shuttleworth, I just don't think this is new news.

I am partial to BSD as a basis for hosting a browser, which is all I want my OS to do. My ideal would be BSD + Chrome mated to be one small, light kick-ass web top. I gave up on Linux a long while ago on the client side. It is chasing Windows, which I think is backward looking.

Keep in mind that BSD underpins Mac OS X.

I disagree

As computers get better I think there could be a market for a really slick compviz style linux desktop with a ton a tools packaged and integrated with it, including virtualbox so you can run your windows xp partition in parallel.

I think there surely is a market for mac osx, even seperated from mac hardware ( if apple would allow that to happen ), likewise there is a market for a linux desktop, we just haven't seen one good enough yet. But we are close.

given the problems with Vista and the success of Macs with OS X I expect some of the other players (Dell, etc) to start looking at investing money in Linux development so they have a compelling OS to ship with their machines, which tells me there should be plenty of opportunities to make money from a Linux desktop.

If you remove the last word from that sentence, you get my opinion.

Ubuntu may try exclusive tie-ups with OEMs like Microsoft does.

20XX is the year of the linux desktop!

if $current_year == $current_year + 1 then:

now is the year of linux

He's talking about selling Linux OS in a shrinkwrap package. On the other hand, Shuttleworth says that services are where the money is at -- something that companies like Red Hat have known for a long time.

Shuttleworth's most interesting comment was a suggestion that XP is being given away for free to certain OEMs. That would be a bold move by Microsoft. The whole ecosystem could change if Microsoft changed XP over to the Firefox model of making money from software.

Yes, MS already sell XP Home OEM for $40 for the eeePC... they have the capacity for brutality. From sales so far, it doesn't seem to need to be any cheaper (in fact, a higher price would probably be OK). The ecosystem built on XP's platform is still worth a lot.

What could kill them is a crossover to the web as a platform... ie. as the desktop.

Implementations of Flash, Javascript and Java are getting faster and adding features needed for the desktop (e.g. client-side persistence). It seems inevitable that the old settop box dream will be realized. It was such a cool idea!

The OS might even be non-linux, if it has other features that are needed more. e.g. faster, more power efficient, maybe intrinsically multicore, so it can work on cheaper, lighter devices, requiring less power - like a phone. Linux has bloated, through evolving in the same rich hardware environment as Windows.

> Linux has bloated, through evolving in the same rich hardware environment as Windows.

I'm not disputing your point directly, but I'd be interested in your data.

For example, there are Linux distros which run well on what I consider low-power devices (I'm thinking of ARM-powered NSLU-2 or WRT54 devices) with, say, 32Mbyte RAM.

I'm also not sure what you mean by 'intrinsically multicore' - are you implying that Linux doesn't make use of multiple cores? Or doesn't do it well?

Linux can be made to be far more lightweight than Windows or OSX, but who uses those distros? Most people use Ubuntu or something similar and those distros are fairly heavy, not as heavy as Vista, but still much heavier than they could be. Things like Compiz and KDE/Gnome really weigh a system down.

> Things like Compiz and KDE/Gnome really weigh a system down.

If you're looking for a modern desktop operating system, then yes, it is tuned for a modern(-ish) desktop.

But if you're talking about using "a different OS" (as the OP was), then you're already further from the mainstream than using a linux distro tuned for your needs.

I'm just checking on the substance of the "Linux is bloated" claim, and trying to find out what alternatives there might be.

Also, I believe that the kernel has also grown much larger, especially in recent years (don't have data or links, just from articles I've read).

Concurrency isn't really a solved problem. If someone did solve it, and formed an OS around it, that would be what I mean by intrinsically multi-core. It might also involve an entirely different CPU architecture, and view of the hardware, which the fundamental concepts of linux just didn't map to efficiently (e.g. no filesystem). Hypothetically speaking. But if you could make it look the same to webapps, it wouldn't matter.

My point isn't that linux is bad in any way, but that if your platform is the web, then the OS doesn't matter. BTW: I also don't believe that linux is the ultimate OS - it's just an OS, and it's possible to have something better. In practice, the alternative might be based on linux (or at least compatible). Not because linux is perfect, but it's there already and because of the ecosystem of software that relies on it (borrowed from unix). Linux is a market success - by which I mean many people use it.

The only alternatives I can think off the top of my head are: BSD, solaris, minix, QNX, Plan 9, AmigaOS and there are also some really "weird" OS's in academia, many no longer used.

I believe Plan 9 was written by the creators of unix... as an improvement. But it lacked market success.

EDIT BTW I'm not claiming that linux is as bloated as Windows - just that it has increased significantly in size, and due to the same cause: there is no real need not to.

> Also, I believe that the kernel has also grown much larger, especially in recent years (don't have data or links, just from articles I've read).

OK. The kernel runs happily on a 32Mbyte ARM system I have to hand. dmesg tells me the kernel usage (a 2.6.21) at boot time is ~2Mbytes:

<5>Memory: 30292KB available (1912K code, 153K data, 84K init)

I don't think that justifies being called 'much larger', unless you're comparing with mid-90s kernel versions or something.

The source code for Linux is much larger, but I think that's mostly due to broader support for hardware and other optional components.

That's pretty impressive, though I was thinking from the beginning, and "recent years" being the last 10 years, and also of the kernel size (not its RAM usage). I agree that kernel size matters less than RAM usage.

But to share my perspective: My first computer was a ZX81, with 1K RAM total (i.e. 1024 bytes). 32*24=768 bytes were used for video in the worst case (it was compressed, newline style). To be fair, it didn't have a filesystem (no disks) or multitasking. In truth, I found 1K just too limiting, and went for the 16K rampack addon.

Regarding my main point: I doubt the massive software base can be given up at this point (even the disruptive technology of cell phones switch to linux when they get powerful enough) - it's just my yearning for those days of daring frugality, when a byte was worth fighting for.

> it's just my yearning for those days of daring frugality, when a byte was worth fighting for.

Which I certainly share :-) I borrowed a ZX-80&ZX-81 and then got my own Spectrum...

Perhaps buy yourself an NSLU-2 http://www.nslu2-linux.org/, they're pretty cheap, low power devices with 1 ethernet and 2 usb ports. There's good support for plugging lots of things into the usb ports (you can stick a usb hard drive or flash stick in for storage, usb sound card for media player, printer server, etc). Just don't expect them to do too much (the CPU isn't powerful).

I use one as a low-power 'always on' file/ssh/git server.

And you can then once more have fun writing software which will run well in a restricted memory environment :-)

My main PC is currently an eeePC, which is not all that far from that! I predict another generation which is even lighter, lower wattage and cheaper.

Those dime-sized java devices appealed to me - can't find the one I mean, but here's similar: http://www.jopdesign.com/cyclone/index.jsp and http://wireless.sensorsmag.com/sensorswireless/Wireless+News...

WARNING: RAVE

It's the simplicity of the ZX81 that appeals to me - especially the small number of commands available, in BASIC and Z-80 machine code. It's concrete and graspable, and possible to get on top of to master some of it perfectly - a toy world, a tiny playground to play in. The small universe makes it easier to be sure you've got the best solution, and limited options make choices clearer and easier to evaluate.

I've managed to get some of that feeling from Java and discrete mathematics - though that essential thrill of efficiency and certainty is missing...

In my current project, I'm trying to create that kind of playground for users, where the choices are concrete and combine orthogonally. I'm aiming for it to appeal to "my people" (i.e. who appreciate what I do), instead of to the naysayers, or opinions about how you "should" do things.

I think it requires a kind of genius, to make the abstract concrete. To me, it's a worthy goal. Fortunately, you can create a work of genius without actually being a genius - provided that (1) you spend a lot of time and effort; (2) you make definite progress; and (3) can link those pieces of progress together, so you build higher and higher. Well, I sure hope I can. :-)

This is a bold-faced lie. I use fluxbox, and honestly, I have never even used the swap partition on my computer with 512mb of RAM. I don't even recall seeing it use over half of the system memory.

Wtf? Are you being sarcastic? Your standard Ubuntu install doesn't use Fluxbox, it uses Gnome+Compiz.

services are where the money is at

In that case, it's still very hard to make money from the Linux desktop. The consumer desktop has no services (except maybe the Geek Squad) and service/support revenue for corporate desktops is probably even smaller than for servers.

It's a service when Dell comes to you and says, "How's about we give you a pile of cash and you make Ubuntu work with our new netbook?"

Porting and certification? I wouldn't want to rely on that as a business model either.

Even if it were free, I think it's a sub-optimal solution. I think they just want to keep pc associated with windows.

As an everything you need is in the box solution, the linux Eee with ff/openoffice/ seemed like a good idea. The reports I heard said it worked great & was really easy. Since it's cheap & usually going to be a 2nd computer, it'd have been a smaller risk. Since its an ultra portable, loading fast is a bigger deal then usual.

Cost or not, people still want the XP. MS got away light.

Even though they ahave to go & pirate office for themselves.

Reply on news.ycombinator.com