The Future of Linux is Proprietary

Linux can be made profitable and it can be made so without going the enterprise route or by relying on the traditional services an support model– as long as technology companies are willing to sell the operating system on their own highly optimized and performance enhanced proprietary hardware.

Note 1: All opinions are of course my own, is simply doing me the favor of allowing me to share my thoughts. Thanks given to the OS Emulation Forums at Delphi for their help and feedback, any mistakes are of course my own.

Note 2: At all times the word ‘free’ in this editorial is to mean freedom unless specifically stated otherwise in the text.

Consider for the moment, the example set by Apple. True their Operating System is a proprietary fork of BSD and the MACH kernel, and not Linux, but consider the untold ramifications of their example purely from their position as a POSIX compliant *nix-like Operating System. I highly doubt that in the beginning Apple anticipated the wellspring of open-source and ‘free’ applications that have since flooded the Macintosh platform thanks to Apple’s OS X. The head CEOs may have expected a few universities to port their favorite *nix utilities here and there, may even have anticipated that a few oddball *BSD loyalists would port over a few text editors or some disk utilities; certainly they likely imagined any such port to be command line only. The current reality today with the Macintosh being able to run just about any *BSD\’Linux’ applications you could name (and most of these as GUI applications that run seamlessly blended into the background with Apple’s native applications) would likely have floored them had they known.

Granted, Apple’s Macintoshes are running these applications in a *nix-like operating system, so it should not be so surprising that these programs run on OS X, which is POSIX compliant. Granted this operating system is a proprietary fork of *BSD and the MACH kernel and as such Apple is free to drop and even lock out POSIX compliance should they wish. Granted even that this is all running on Apple’s own proprietary hardware running its exclusive version of the Power PC processor in an undocumented chipset.(Efforts of reverse engineers managing to make booting Linux on these chipsets possible notwithstanding, they are considered undocumented by Apple itself.) Accepting also the fact that Apple is primarily a hardware company and that these desktop computers are running a proprietary Desktop Environment– Granting all these things and many more besides, one fact remains. . .

Apple is ‘selling’ free and open-source applications to the world on their proprietary hardware, and if their quarterlies to date can be believed making a nice profit at it besides.

Now, let’s look at another company making a nice profit from selling open-source software on proprietary hardware, Sharp with its Linux Zarus platform. In comparing the Macintosh and the Zarus platforms, we find them to be alike in many ways. Similar in many ways, but as I’ll explain not quite the same in one important way.

To start with, while its true that the Zarus is also running a *nix-like operating system with POSIX compliance, its kernel and much of its operating system is non-proprietary. Also, while its Desktop Environment IS semi-proprietary and from another company altogether (Qtopia), it runs a customized stripped down kernel optimized for the Zarus device by a team of Sharp engineers who likely know their hardware just as well as Apple’s team knows theirs — if not better due to the PDA’s innate lack of various upgrade hardware cards. Wireless modems and other such accessories notwithstanding, as these are all external to the main hardware of the device itself. Just as Apple has its own applications bundled with each OS X Macintosh, so does Sharp also have a few applications bundled with their Zarus systems. To be sure the types of applications and the capabilities of each system vary, but that is due to the targeted nature of these computers more than anything else. The Zarus has been a widely reported success in its native Japan, each new release garnering multitudes of early adopters. Likely the device, especially in its newest configuration, could be a major success for Sharp in other parts of the world as well if only it were released with competitive pricing and perhaps larger storage ability to show off its impressive capabilities.

Yet, for all their similarities, despite even the fact that both systems can conceivably run the same applications, there is a key difference that will forever set them apart. Apple need not ever worry that someone could one day release without their approval a modified version of OS X to compete with them on the Macintosh, while Sharp must already contend with competition from a distro of Linux called ‘Open Zarus’. To many this seems to be the fatal flaw in comparing Linux systems to *BSD and in the case of ordinary distros I’d agree–that need not be the case here as I’ll try to explain.

If I may digress momentarily, take Lindows for a perfect example. If Lindows makes changes to Linux or to any GNU\open-source software in its distro, then depending upon the license of that particular bit of code and what it may or may not link to, the changes MUST be available upon request, so as to be of benefit to all users. That is commonly understood to mean that if (for example) a Lindows engineer was playing with a bit of code and managed to tweak out a modified library for XMS or some other media player that somehow improved mp3 playing, then later it is released in Lindows version 5.12. When the next version of Xandros, say version 2.2.3 or perhaps a version of Fedora labeled release 1.113 arrives on the scene it will likely also have the modified library, something that results in an improvement to these other distros with little or no investment on part of their engineers. In such a case it is seen that Lindows ‘loses’ even though the improvement they made is now available for the benefit of all Linux users, because the work they did cannot be used to further their brand identity. Even if they could in fact do the same to the others if say Xandros made a release with an updated version of Konqueror that combined the filemanager in XPDE to make an improved Linux eXPerience for newbies or switchers from Microsoft’s Windows Operating Systems. From a purely business standpoint Lindows has ‘lost’ because they put out time and effort for an improvement or feature that loses its exclusiveness before the company can benefit (monetarily) from it.

However in a case where the hardware itself is proprietary on a *nix-like operating system BOTH the company and its users can benefit…

Returning to our two main examples, the Sharp Zarus and the Apple Macintosh running OS X, say that Apple releases the specs of their Quartz Extreme or perhaps their Altivec optimizations and some bright engineer at Yellow Dog Linux notices an outstanding bug here or there, fixes them and submits the patch to Apple before going ahead and improving the performance of Yellow Dog on Apple’s hardware. Perhaps a better example would be someone toiling away at the Darwin core to OS X submitted by Apple to the open-source community who discovers a performance enhancing tweak or fixes an error to file I/O and submits this to the community and Apple picks up on it resulting in an impressive boost to overall OS X performance. Who benefits? Certainly in the first example the various Linux distros on Apple hardware and perhaps Linux would benefit in the second example too. The Open Darwin Project would certainly be boosted by such an event in the second example, since they would be the first to get the improvements to system performance–but within a few days or weeks a service pack would definitely be released to OS X. In the end regardless of which operating system the machine is running Apple benefits as does its customers.. Best of all, despite the Linux and or open-source communities benefiting from these hypothetical developments, Apple doesn’t ‘lose’ anything or sacrifice their brand identity because they have already made their profits selling the machine to their customers and any updates become enhancements that further the value of their brand.

In Sharp’s example what if their Zarus engineers were tinkering around with the source code and discovered a line or two that dramatically boosted the display of embedded Linux as it interacted with the processors used in their devices. They fix the bug or make the improvements and then due to this being in relation to the kernel send their patches to Linus Torvalds, thus the code makes its way into Linux 3.2. Of course it won’t take long before the Open Zarus distro managers find the patch and they have it also. This still doesn’t cost Sharp anything, since they’ve already sold the customer their Zarus hardware. Whether the customer runs Sharp’s own version of Linux or the freely available ‘Open Zarus’ version of Linux hardly makes a difference. Sharp might lose out on a chance to sell operating system updates to a customer who might choose instead to opt for the ‘Open Zarus’ distro since it’s free, but since Sharp has already made its investment from the sale of the device itself such an event would not be seen as harmful to its long term profitability as if Sharp were in the business of making its Zarus Linux for X86 computers. . .

However, by looking at Sharp’s Zarus as example of open-source, free software on vendor specific hardware and projecting Apple’s desktop strategy as a possible model, I think we can see where the future of Linux will be. The X86 desktop is increasingly viewed as an albatross by even its progenitors, even as the ‘standards’ get changed and rewritten at leisure by hardware manufacturers who seem not to care if their hardware works with any platform other than that of Microsoft’s Windows operating system–if even that! Modems and network cards that are hostile to alternative operating systems, fake RAID, and multiple audio\video problems caused by bad or lacking support from manufacturers are all issues forced on us by the X86 architecture. Given things as they are, its a minor miracle in itself that Microsoft has been able to make Windows work as well as it does! In a world that is increasingly shifting its gaze to a 64-bit future, why chain ourselves to a dying architecture? Why not jump ship early and get to work preparing for that 64-bit future? There is no reason why a vendor choosing to create a computer literally made to run Linux, able to run open-source and free applications, and capable of doing it well should be unable to turn a profit.

It will be a brave Linux vendor who first takes a final hard long look at the future and announces their decision to develop their future distro releases exclusively on their own hardware. They will likely in the short term at least lose many of their customers. Or at least likely some who have never bought from them or likely ever booted their distro claim– but if that company is able to stay the course and get past the inevitable missteps along the way they will see that the benefits well outweigh the risks, that their hardships endured and slanders persevered make well the reward worthwhile. As long as they can provide a well performing desktop or other computing device that can run the variously available open-source\shareware (visit Zarus’ online software store for an idea of what is possible!) an do so at a reasonable price they’ll have no problem at all getting customers.

To this end the recent moves by some Linux vendors seem prophetic. Lindows with their email\webstation device. Lycoris with its Linux Tablet device. Both are baby steps towards a profitable future. Shaky and somewhat uncertain, but as the time goes on we’ll be seeing more and more moves like these. Vendors will be providing specific hardware with which to run their Linux distros and open-source software, us better performing computing devices and the companies themselves will have a more profitable Linux.


  1. 2004-01-09 8:24 pm
  2. 2004-01-09 8:48 pm
  3. 2004-01-09 8:48 pm
  4. 2004-01-09 8:49 pm
  5. 2004-01-09 8:49 pm
  6. 2004-01-09 8:53 pm
  7. 2004-01-09 8:53 pm
  8. 2004-01-09 8:56 pm
  9. 2004-01-09 8:57 pm
  10. 2004-01-09 8:58 pm
  11. 2004-01-09 9:01 pm
  12. 2004-01-09 9:03 pm
  13. 2004-01-09 9:14 pm
  14. 2004-01-09 9:15 pm
  15. 2004-01-09 9:18 pm
  16. 2004-01-09 9:24 pm
  17. 2004-01-09 9:26 pm
  18. 2004-01-09 9:32 pm
  19. 2004-01-09 9:33 pm
  20. 2004-01-09 9:34 pm
  21. 2004-01-09 9:36 pm
  22. 2004-01-09 9:44 pm
  23. 2004-01-09 9:54 pm
  24. 2004-01-09 10:05 pm
  25. 2004-01-09 10:05 pm
  26. 2004-01-09 10:21 pm
  27. 2004-01-09 10:24 pm
  28. 2004-01-09 10:43 pm
  29. 2004-01-09 10:46 pm
  30. 2004-01-09 10:50 pm
  31. 2004-01-09 11:07 pm
  32. 2004-01-09 11:20 pm
  33. 2004-01-09 11:58 pm
  34. 2004-01-10 12:04 am
  35. 2004-01-10 12:13 am
  36. 2004-01-10 12:27 am
  37. 2004-01-10 12:31 am
  38. 2004-01-10 12:34 am
  39. 2004-01-10 12:38 am
  40. 2004-01-10 12:49 am
  41. 2004-01-10 1:47 am
  42. 2004-01-10 2:34 am
  43. 2004-01-10 2:46 am
  44. 2004-01-10 2:49 am
  45. 2004-01-10 4:53 am
  46. 2004-01-10 6:32 am
  47. 2004-01-10 6:37 am
  48. 2004-01-10 7:01 am
  49. 2004-01-10 7:03 am
  50. 2004-01-10 7:14 am
  51. 2004-01-10 7:26 am
  52. 2004-01-10 8:16 am
  53. 2004-01-10 9:09 am
  54. 2004-01-10 9:10 am
  55. 2004-01-10 11:14 am
  56. 2004-01-10 2:52 pm
  57. 2004-01-10 5:45 pm
  58. 2004-01-10 9:45 pm
  59. 2004-01-11 3:56 am
  60. 2004-01-11 11:15 am
  61. 2004-01-11 10:21 pm
  62. 2004-01-12 1:51 am
  63. 2004-01-12 3:32 am
  64. 2004-01-12 2:51 pm
  65. 2004-01-12 5:16 pm
  66. 2004-01-13 11:57 pm