If eWeeks Jim Rapoza is right in his most recent piece on securing Windows then only 10% of users have converted to Windows XP. That seems like a low percentage (although from working assignments I know a lot of shops that are still mixed NT, 9x, and 2000 operations. So there may be some substance here). But say Jim is off by a factor of 3 – that still leaves a sad story for Microsoft.
So if Jim * 3 =30% XP penetration is right, then the absolute nature of the 2006 deadline for the next generation of Windows as WhateverHorn is more understandable. First, Microsofts Software Assurance Plan holders want more than Windows XP Service Pack 2. They have paid for 3 -4 years of software with an expected upgrade which was supposed to include a path from from NT/9x/2000 here to a much more secure, reliable, and functional WhateverHorn there. In short a compelling upgrade and not in 2007 or 2008 but it should have been happening now.
Quick aside. Back in 2002-2003 when all those Software Assurance deals were being signed Microsoft was promising Longhorn by the end of 2004. In May of 2003 Sr VP of Windows Will Poole promised Longhorn coming to market in 2005. (See page 17 of the August 06 2004 issue of Infoworld for the detailed timeline). In March of 2004 Bill Gates concedes that speculation about Longhorn making 2006 “is probably valid”. In August 2004, after major revisions and downsizing/resizing of Longhorn (it is still quite speculative as to what is in or out of the emerging OS package), a firm 2006 delivery date was established.
So now IT shops have to deal with the fact they had committed to Windows based on the premise that they would get an operating system by 2004-no, 2005-no, 2006-not really likely. The question outstanding is whether the current great uncertainty about what and when the new Windows appears – is that good enough ? And is the WingITWindows going to provide a compelling story in the Windows weak areas of security, interoperability, reliability and scalability. And is the associated development environ, the Visual Studio/.NET Framework going to be less fragmented than the current version before, during and after delivery of WhateverHorn. And in the area of the functionality there is grave concern whether Redmonds Brobadingnangian desktop vision of more more more – consuming the state of the art in CPU power, GUI complexity and system resources – does this really fit the agile, interoperable, and scaled-to-fit world of network and many device computing ? And has the Black Widow Spider of 1 Microsoft Way bitten and poisoned too many OEM and ISV partners to work co-operatively in these rapidly evolving markets?
So every Linux pundit worth his Penguin suit is crying “great opportunity”!
But this party is not so sanguine. The evidence of the diversity of the desktop market would seem to point to mixed blessings at best. Lets follow the logic.
The Windows desktop market can be broken down into at least four major battleground segments:
1)the corporate IT segment that values the “-ties”: security, reliability, availability, usability, functionality, manageability, interoperability, scalability, flexibility in roughly this order(Note these “-ties” are in a different rank order for a server OS). If Jim Rapoza is right about XPs adoption rate it is surely the conservatism of corporates that is holding XP back;
2)the government IT segment that is so cashed strapped and being demanded to provide better services that they can ill afford the new high-cost desktop provider that is Microsoft. Hence the revolts in Israel, France, Munich, Florida and dozens of government buyers (city, state/province/region, and national) throughout the US and the whole world;
3)the underdeveloped and just emerging markets such as many parts of China, many parts of Africa, Southeast Asia, etc. These are different from the government in that most do not have at least some computing power on the desktop(dont get me wrong there are tens of millions of desktop computers in China, Asia, and other developing regions. But if household and business penetration is 1/4 of the developed countries I am prepared to be surprised) – and the training and savvy that implies. But the lesser developeds do often have the same cash/budgetary binds as governments;
4)the developed countries in the Home markets. Although PC computing is stuck at about 2/3 penetration of households in developed countries, this is a 90%++ Windows PC world.
Now if we examine each one of these segments for the Linux versus Windows advantages, then we get a better assessment of Linux Opportunity.
Lets take the easy analysis first. Segments 2) and 3) are most fertile for Linux but for different reasons. China and many other developing regions want to establish their own security, local custom device functionality, and local software providers first and foremost – and Linux provides the Open Source starting point. But China Linux, Africa Linux or Indonesia Linux will be very different from Suse or RedHat especially on the desktop. But the common key is that these Linux will be homegrown. Expect a strong and dominant HomeGrown Linux presence except in countries like India and parts of China, Singapore, Malaysia where outsourcing to the US and Europe will mean a must have Windows presence as well.
Cost, security and responsiveness to local customization will be the driving factor for segment 2)governments – and unless Microsft really sweet-deals and locks tight legally for many years as in the case of England, this market presents the biggest opportunity for Linux VARs and providers. Now we get to see if those Microsoft TCO stories are true or false – or conversely whether the Linux vendors and VARs are well organized or not.
The developed corporate IT segment would seem to present the same opportunity as governments but for one factor – desktop power users from engineering to finance to operations to marketing and on and on still depend on the great Microsoft advantage: all those tens of thousands of Simul8, AutoCAD, Maximal Optimizer, @Risk, and other Windows-only apps. And on the petal to the metal, thin client data entry apps Windows Terminal Services + Citrix have kept thin clients in the Windows fold. Sun, Wyse and others have beat their heads on the thin client market and have been effectively rebuffed. Why should Linux vendors do any better?
Well to just shoot down my own theory – Linux vendors are doing better, but in small markets and lead by a software app that takes advantage of Linux and other Open Source and then adds proprietary service , utilities and support (usually all three).
But this is hard slogging. Dont expect expect a Rene Thom-like Butterfly catastrophic fall in Windows market share. Hell, IE is a clearly inferior browser and yet it has managed to gain market share until recently. But do expect a gradual or even occasional precipitate fall off of Windows share as proprietary, only-WebServices interoperable, and continually security/reliability/availability challenged Windows loses out like in 3D rendering workstations or large scale manufaturing simulations and other one-by-one markets to Linux.
So this leaves the Home market. As another aside let me say that my current 25% usage of Linux would jump to 100% if Adobe Photoshop or Jasc Paint Shop Pro and Macromedia Dreamweaver were offered on Linux. But Apple Mac OS/X users will say they have a better Linux with most of those desired Apps and then some. However, in response I will say that Mac has always been the a) the premium price spread in computing and b)adopted an even more ruthless scorched-earth policy than Chairman Bill on backward compatibility and interoperability. And really does such “single-minded greatness” that produces a system with a mouse which has only one button deserve such slavish loyalty ?
As you can see from my aside, the Home market is filled with semi-fanatic irrationality. But if I had my druthers I would ask you which vendor, Linux or Windows, is better poised to take advantage of the coming big trends in Home living ? HDTV and the improved sound systems coming over the airwaves. Wanting to manage and interchange readily the whole audio, video, TV , telephone/messaging/VOIP experience ? Which OS is poised to take advantage of Speech and more importantly command recognition – a much more limited and therefore reliable recognition problem. Instead of Clapper its “Lights Off!”, “Heat Up”, “Ice in a Cup” , etc. I know Microsoft has killed off/assassinated 4 times as many partners than all the Linux companies combined have reached an agreement with. And thats the whole problem.
In sum expect, Microsoft and Windows (even the bloated LongGongNowhereHorn oops …. I promised never to do this again) to do very well on its home and starting turf. There could be some big Linux/Java surprises on specialty devices, mobile phones, PDAs. But dear readers, tell me how is Linux going to be able to displace Windows from being the premier or the controlling Home desktop ???
As for Servers – now thats a completely different story.