Should Linux Play the Apple Card?

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

Periodically, the suggestion is made. The latest similar argument – though he does not, in fact, take Apple’s name in vain – comes from Jeremiah Foster in the form of an open letter to Mark Shuttleworth, of Canonical/Ubuntu fame (oh yeah, and Soyuz).

The missive makes the same case I’ve seen and heard elsewhere, though certainly not in this direct a form. Distilled down, the argument runs as follows:

  1. Supporting an operating system on all hardware platforms is hard
  2. Supporting an operating system on one platform is easy (relatively speaking)
  3. Proponents of desktop Linux, therefore, should do the latter (though obviously not at the expense of the former)

In other words, desktop Linux advocates should borrow a card from Apple, whose job is made infinitely easier by virtue of the fact that they only officially support one hardware platform – theirs (and even then they have random shutdowns, melting power cords, and machines running hot enough to fry eggs on).

When I’m asked whether or not I agree with that strategy, I usually hesitate. The short answer is, yes, I think that would pay dividends. It’s absurd to me, as an example, that it’s 2007 and my latest laptop (Thinkpad X60s) suspends to RAM less reliably under Linux than did my last one (Thinkpad X40), although apportioning the responsibility for that particular situation would likely be an exercise in finger-pointing. But it’s difficult to argue that a simplified (and rigorously tested) hardware platform would not yield performance and reliability benefits; that seems obvious to me.

But the long answer is actually: maybe. Not that I see any harm in a software maker and hardware maker trying it – quite the contrary, I’ve recommended as much to both sides of that equation in public and private. That said, however, I think invoking the “Apple card” gives people unreasonable expectations of the kind of integration that could be expected from such a partnership.

Let’s say for the sake of argument that you could get a Canonical to partner with an Acer, as Foster recommends, or perhaps a Dell or Lenovo. That’s all good, and you at least could ensure that users at least had the correct settings for X out of the box. You might, with some work, even be able to get suspend working more reliably than Windows (XP comes back about as frequently as Linux on my current hardware). Noble ambitions both.

But where are you going to get your graphics cards? Last I checked, most of those come from either ATI, Intel, or nVIDIA, and with the exception of Intel they don’t happen to be particularly Linux friendly. And how about sound cards? Or even more critically for certain types of users, embedded wifi/EVDO/UMTS/HSDPA cards ? And let’s not even get into the iPod?

Before the Linux advocates yell “gotcha” and note that every single one of those devices can, in fact, work on Linux, let me say: I know. I’ve had every single one (with the exception of EVDO, which I’ll probably try to get running this weekend) working at one time or another. Most of the above works – to some degree – on my current machine.

None work perfectly, however, and virtually none of them worked out of the box. I’m willing to put up with that in return for the benefits I get working off a Linux desktop; most, however, would not be. The point, as is probably obvious, is that a single partnership here is not likely to suffice. What’s required instead are different hardware vendors perceiving the opportunity that does exist – go see the Dell Ideastorm site to get an idea at least for the enthusiasts market – and collaboratively working together not on an exclusive platform, a la Apple, but a more Linux friendly one. Hand choose the components that are known to work best with the platform, and build from there. It may not get Linux to Apple’s level of just works, but it can’t hurt either.

But let’s acknowledge along the way that it’ll take more than just Canonical and your hardware manufacturer of choice; it’ll require firm commitments and support from a variety of hardware suppliers. Support that’s been difficult to come by to date, despite the fact that it would be a very significant differentiator.

Hmm…maybe it’s time for me to write an open letter to Canonical/Intel/Lenovo 😉


  1. But this ISN’T the Apple card. IN fact, unless Ubuntu starts making hardware, that’s not a card they can play. Apple doesn’t simply support one set of hardware, they make it. Could Ubuntu choose to support just, say, lenovo? Sure… but that does not give them control over the hardware. They might be able to work a partnership where they get pre-release hardware and can thus test on it before it’s in the wild. But what if they find that an issue exists which can only really be fixed by a hardware or firmware rev – will Lenovo do that?

    An alternative would be to do a very tight partnership with a small hardware vendor that WOULD work in a tightly coupled manner with Ubuntu (or some other distro). But then you have to market that and get it on the approval list of companies (if that’s a market you want) and back up the issues that still arise with support… because if you’re going to tell me to choose a particular hardware vendor if I want to make sure Ubuntu will run well and you don’t offer support – what have I really gained?

  2. *Something went wrong, I couldn’t post correctly the first time around*

    You’ve just missed the point of opensource itself. Its about choice. By restricting support to a particular platform, you reduce the options for the end-user.

    So no, I don’t agree playing the “Apple card”.

    Instead, there should be third-parties that sell ThinkPads (as example) with pre-installed Linux, as well as provide support for it.

    The hardware companies just make the hardware and provide the necessary details for hardware support in Linux. Whether it be an open driver developed under NDA, or supporting an open project that is developing the necessary driver.

    The community provides the software, and the third-party companies bring these both together for the end-user, as well as providing support and custom patches for the specific hardware.

    …And this is EXACTLY what’s happening. I suggest you visit companies like:

    (International delivery)

    (US only)

    The objective for Apple, is to sell the hardware or platform as a whole. Their OSX is only a complementary part of their system. Their idea is to have full and total control of the solution. (its a completely different goal, and kinda contradicts the intent of opensource goals).

    As for video card drivers…Open ones are being worked on. Nvidia card owners will have the Nouveau project, while Ati/AMD video card owners will have the R300 and the upcoming R500 project.

    And the power saving issues? Blame that on the manufacturers for sticking to primitive BIOS firmware. They should assist and adopt (or at least be compatible) with LinuxBIOS project. (I only see Tyan supporting it in most of their mobos).

    This results in systems booting much faster (3 to 5 seconds) and dumping all the powersaving functions onto Linux itself, instead of Linux needing to understand the crappy proprietary BIOS powersaving features and function calls.

    Overall. I don’t agree with your opinion. You suggest the “quick and easy way” approach. (short term gains at the cost of opensource ideals)

    I suggest the long term approach in attacking the problem. (Tedious and a test of patience, but doesn’t compromise the basic principles of opensource itself.)

  3. Open source is about source that is open, not choice. Choice, for both users and developers, is a consequence of openness.

    It’s obviously easier to develop better software if you control the hardware on which it runs. A relationship with a hardware vendor might provide that kind of control. But, I doubt any vendor would want to participate unless the software ran only on their hardware.

  4. Personally, with all its geniouses and $ in the bank, Google should build a nano-ITX box (http://www.linuxdevices.com/files/misc/nano-itx-board-big.jpg) with a reworked Linux (GoogleOS) that would be sold online and through retailers, and sport a disk drive: GoogleOS would only allow downloading applications from Google’s software library, ensuring that applications work right out of the box. Needless to say, those boxes would also self-update every day, keeping them safe.

    Most people don’t care about how their computer works. They just want to use a few, standard applications. At this point, there just isn’t any solution for those people. It just takes one virus to ruin their computer and have them go back to the store to have XP reinstalled from scratch.

    Others who want total freedom can do it on their own.

  5. rick: you’re correct, and i should have note the distinction. nor did i mean that Ubuntu should be supported on a single hardware platform – that would be suicidal.

    i meant, instead, what you suggested; tightly partner with an existing supplier to deliver an offering optimized for the Linux experience.

    fwq2_lee: nothing wrong, precisely, you just got moderated – but here’s your comment.

    anyhow, respectfully disagree that i’ve “missed the point of open source.” i’m not talking about an exclusive arrangement, merely a further, more specialized partnership. no reduction in choice, just a special focus on one.

    billg: i get that it’s about choice, but the point is that we have that now and it works suboptimally. so how do you remedy that without negatively impacting available choice? tight partnerships are one option.

    as for being exclusive, i don’t think that’s necessary. if Lenovo, say, worked with Ubuntu and Intel on the work that combination would itself be fairly unique without being exclusive.

    Vincent: agree that most folks don’t care about administering their machine, but disagree that Google should maintain its own library of apps. we have too many of those already; one more would be needlessly redundant.

  6. “Should Linux…”

    So, there is an acting entity called Linux who is or should be sorry about “most people” who, thanks to the lack of skill, have problems as in

    “I’m willing to put up with that in return for the benefits I get working off a Linux desktop; most, however, would not be. ”

    or worse, aggravated by the lack of initiative, as in

    “Most people don’t care about how their computer works. They just want to use a few, standard applications. At this point, there just isn’t any solution for those people.”

    I cannot understand that kind of thinking. Even now, Linux is a cooperative effort. People who do not contribute to Linux are welcome to get the benefits anyway, but they cannot have any say. That is, if you do not have skills necessary to contribute (cannot read man and configure, cannot compile, etc.), then shut up and pay for Windows.

    There is and always will be a limit to what one can get for free. If you still want Linux and do not know how to configure your hardware – buy a distro with tech support and call it, period.

    There is no “Linux” to play the Apple card. There are hardware vendors and distributions vendors. Let them think if it is time to cooperate and preinstall a distro of one vendor on the hardware of another vendor. As far as I know, Dell will do exactly that as soon as it makes business sense.

  7. This is being done over at Emperor Linux but Lincoln & Adriane Durey.

    Their value-add is not trivial and not free. But, there you are.

    Excellent place to get a machine!

  8. That’s why we’ve started the Open OEM Project. We are trying to figure out the best way to build a computer that’s truly open, has open source software pre-installed and “just works”. Now I know it’s easier said that done, but I think we all need to support this project and make it a reality in the near future.


  9. I was writing about this same thing on a /. post yesterday. Dell could do something like this by working with (for example) Ubuntu on Dell-buntu for their products. Tightly control (at least at the outset) the hardware available and set up a robust user forum site.

    Dell’s commitment to service could be that they would provide drivers and configuration directions based on whichever system you’ve got, list products that are agreeable with their Linux configuration, and in the process become the darling of Open Source folks everywhere.

  10. […] (among other things) – could use a sexy platform over which they have total hardward control (AKA the Apple card). And for the folks that don’t want Solaris, well as it turns out Sun OEMs Windows now. Or […]

  11. […] Should Linux Play the Apple Card?, RedMonk – tecosystems, Stephen O’Grady (Blog) Permalink |  Email This Post | Technorati Links | Bookmark on del.icio.us | digg it Categories: Links […]

  12. […] than it is on Windows. No, I don’t expect Google to tackle the very difficult question of device compatibility at all in the initial releases of Chrome OS, if only because it’s a slippery […]

Leave a Reply

Your email address will not be published. Required fields are marked *