Infoworld, which normally is pretty IT prescient, has a TechWatch article from November 2003 in which they declare “Next year  just may be the year that desktop Linux emerges as a viable alternative to Windows.” The drivers are a continuing stream of Windows desktop security and reliability problems (check – still going on); ever increasing prices for Windows software and support (check – having created the security, spam and phishing problems; Microsoft is now going to charge users for maybe fixing some of its malaware problems); and the ever receding time horizon for the Longhorn BigFix (check – even with dropped WinFS and other features, Longhorn will be hard pressed to make a mid 2007 intro).
So all the conditions are ripe for the Linux Desktop takeoff. And as the article pointed out the PC stars such as Dell, HP, IBM (Lenovo aided), Novell, and RedHat among others were (and maybe still are) ready to pitch. And not just hardware+software but also support programs particularly in the corporate arena.
But it aint happening.
Why ? Because “Linux” is already here. Two players which have Linux-look alike capabilities are already in the market and they are doing respectively pretty well and “just kicking the tires but you guys used to have a great workstation product”. I am talking about Apple OS/X and Sun Solaris 10 respectively. Both may have a better chance of capturing some serious Windows desktop real estate than Linux.
Linux is plagued by its constitution – GPL. This allows anybody to distribute the product and even now, from time to time, just about anybody does. True, there are now about 4 major Linux distributions RedHat, Suse/Novell, Debian, Mandrake but the corporate IT community has already voted and they really want ONE. And so RedHat leads the Linux Parade by a fairly wide margin despite having cost+support prices within horseshoe distance of Windows. The IT community apparently craves oneness in operating systems – one supplier to bind them all, one supplier to rule them all, one supplier to make all IT shops rue that one supplier decision.
So Linux on the desktop is really competing against all the various distributions plus two incumbents that are already there on the desktop and that have many Linux like attributes. So the horse race is not among just the Linux distributions but all that espouse the Linux Pedigree.
The Linux Pedigree that OS/X and Solaris 10 aspire to are secure, reliable, available, compatible, robust functionality, open and a good plus growing selection of applications. Oh and cheap too. Lets see how the two pretenders to the Linux crown measure up.
OS/X has a pretty good record on security helped by the fact that Microsoft dropped support of IE on Mac and Safari has Gecko roots and therefore reliability and security. OS/X and reliability have a mixed history depending on who you talk to; but certainly a checkered backward compatibility story. But another proxy for reliability – availability would appear to give the current editions of OS/X a pretty good record. And of course Steve Jobs would argue that the the number of unique and interesting applications is both growing and a part of the Apple heritage. But Steve would also argue that OS/X is cheep-cheep-cheep at double the price. But dont try to open the door on Open.
Solaris X has all the secure, reliable, available, backward compatible and industrial strength virtues stretching back to the days when it was a workstation powerhouse. And those virtues are redoubled in Solaris 10 with multiple CPU, 64-bit, grid-enabled computing. Hidden virtue – Solaris 10 supports two hardware architectures close to equally. This is a feat no other OS but Linux can come close to claiming (Windows and StrongArm for the Pocket PC … yeah, right!). And Sun is vowing to make Solaris open but it may be just the RedHat leger-de-main – support prices will kick it back up to Microsoft levels. We should know by Valentines Day (give Sun a few weeks after-launch time). But the critical problem for Solaris 10 is the collection of applications and their growth rate on Solaris. Ever since Redmond kicked the bejesus out of Sun and SGI and HP and all other Unix workstations the number of non-vertical apps on Solaris has, as a Solaro-phile would say, not mushroomed. Will open and very good and free mean more good Solaris apps ? Will Solaris 10 slip on robust and reliable because of a rushed Jan 30 2005 deadline ? Stay tuned.
So Linux on the desktop has not taken off because there are some Linux lookalikes already in the market and the market craves Oneness and now with IT operating resources tighter than a witches …….. ZAP! Yikes![I didnt write that!]……. the inclination to switch may be there. But where and how and how much dollar and time and people resource pain to be endured to accomplish the transition is the question. One or more of those ingredients may be in serious short supply. So 2005 will be the year to watch for signs of how the OS Desktop change winds will blow.
First shops will have to come to grips with 64-bit computing on the desktop in 2005 especially for compute bound workstations: development, database and system support, graphics and video intensive computing, industrial design and simulation etc. Next dual and multi-core computing with virtual machine emulations will also go critical in 2005 – again requiring a Windows upgrade. Finally, the home media centre is due to happen anytime soon just like LANs all during the late 1980s. Again in 2005 the question will get asked – are we there yet ? These are just the preliminaries for the 2006/2007 tsunami that is Longhorn.
So place your bets for the OS Desktop Horse Race of 2005 – will it be Evil Empire, Open Sourcery, iApple or that dark horse, Solaris 10????