Cheap PCs are not Microsoft's friend

Three interesting trends in the world of netbooks. First, they are estimated to account for about 8% of worldwide PC demand, climbing to 12% over the next four years. That's an astounding number. Second, 30% of netbooks are believed to be running some version of Linux, which equates to 2.4% of the world market. Another astounding number. Third, the remainder only have the capability to run a stripped down version of Vista, which will have a negative impact on Microsoft's revenue.

"It’s pretty clear that netbook sales are cannibalizing sales of higher priced versions of Windows," said Matt Rosoff, an analyst at Seattle-based research firm Directions on Microsoft. "This is clearly a trend Microsoft is concerned about."

I don't know where Mr. Rosoff gets this information, but it's hard to argue with his conclusion.

Consider who would be the primary target market for these devises. Students. Middle-market and down-market consumers. Bargain hunters. In current conditions, I believe that netbooks will make market share gains even greater than the predictions indicated above. On the other side of this recession, Microsoft may find it difficult to win back the consumers that have explored beyond the fence. Further, it will be imperative for organizations to build customer-facing applications to accommodate Linux users.

More information from the Wall Street Journal: Cheap PCs Weigh on Microsoft

Virtualization returns to the desktop

The Wall Street Journal has reported that "IBM says it has created a 'Microsoft-free' virtual desktop -- a complete suite of applications that run on a backroom server and don't require Microsoft software or costly desktop hardware." The basic idea is to deploy desktops with virtually no locally installed software. All of the tools the user needs are installed on a server which reserves memory and processing power for each logged in user. The user experience is similar to a conventional desktop, including the ability to store preferences. However, the desktop PC can be a very low-end machine.

A few years ago, I had set up one of my clients with Linux desktops and a terminal server connection to a Windows platform. Users had all of the benefits of Microsoft applications, like familiarity and compatibility, but the company could easily get by with sub-$500 PC's, including the operating system. We shifted the budget in favor of high-end screens, which thrilled the users. More sophisticated users took advantage of the locally installed Linux O/S to install cookies from websites (which we restricted on the server-side).

Kyle Vickers, CIO of the American HealthCare Association, has long used a Citrix environment with dumb terminals on the desktop with great success. His team has no software on the vast majority of user desktops.

So IBM isn't -- by far -- the first to come to this party.

There is, however, one very simple reason why this has not gotten widespread use: Portability.
The virtual desktop simply does not yet work for the mobile user. It's coming, for sure, but we're not quite there yet. Until the executive user can connect her laptop to the corporate server from an airplane, this extreme desktop virtualization will remain a true rarity.

WSJ Article: IBM Creates 'Microsoft-Free' Desktop

A path forward for Linux

One of the big issues for Linux, in general, is that the variety of distributions makes broad-scale adoption of specialty applications relatively difficult. The general use applications - like web browsers, word processing, etc. - aren't impacted because developers have taken the initiative to adapt these apps to the various applications. But business lives and dies by the usability of their specialized software, be they manufacturing control, database or order fulfillment systems. Until some order is established in the Linux distribution community, this problem will persist.

A writer at InformationWeek proposes several improvement strategies, including some sort of governing body to oversee package development, among other things. Most of the suggestions are sound, but I believe the proposal is somewhat idealistic, if not naive. Nevertheless, here they are, and you may notice a couple of themes ...

1. Meta-package format: a promising potential solution because I believe entrepeneurs could create a sustainable business out of provision and maintenance.
2. A consistent configuration system throughout (from the kernel to userland tools and user applications): a great idea, but who will decide what the standards are to be?
3. Consistency in the kernel application binary interfaces: I don't pretend to know what this means, but the problem with deciding who sets the standard presents itself yet again.
4. Native File Versioning: This will make it feasible to consistently roll back installations and upgrades. Good quality software is built with this in mind, and shouldn't be too much to ask of the developer community as a whole.
5. Consistency across Audio APIs: It seems that every distribution has its own "best" audio implementation. Standards would be nice, but again... who?
6. A predictable GUI: Every distribution has a different look and feel, which confuses most users.
7. X11/Application Integration: Another hallmark of good software is its ability to gracefully handle errors and hiccups. But many Linux-capable applications lack an ability to restore a previous session.
   

Consistency and Standards.
One of the great things about dictatorships is that standards are set, and consistency is enforced. Microsoft has proven this and, as mediocre as some of their offerings have been, it has been an unimaginably fantastic benefit that developers and end-users know what to expect when they sit down at any Windows PC. Of course, they profit from this benefit -- as they should. And therein lies the path for Linux.

Someone needs to be able to make ridiculous profits from standardizing the Linux offerings. Until that happens, Windows will dominate the desktop.

InformationWeek Article: Fixing Linux: What's Broken And What To Do About It

4G - Speed to burn

I have had a unique opportunity to test Sprint's nascent 4G WiMax network. Thus far, the results are astounding. Several speed tests have clocked in at 8Mbs download and nearly 3Mbs upload. Coverage is very limited at present, as Sprint builds out the network, and there are - as one would expect - very few users using the current bandwidth. However, the potential can't be overstated and I would keep an eye out for its general release in your area.