The Incredible Shrinking Operating System

Continuing in the theme of discovering goodies on the Web whilst otherwise enjoying downtime over the holidays (still underway here at Farpoint Group, where progress on our many holiday projects is otherwise at best mixed), consider this tidbit from InfoWorld. I have long believed that operating systems, at least on the client side, have outlived their usefulness. As this technology has become productized, it's become ripe for bloat, as we've seen (only for example) in just about everything Microsoft has produced. OS X is pretty good, but not exactly transparent or as easy to use as many claim, and LINUX is but the latest incarnation of UNIX, which can trace its roots back 40 years and has consequently also suffered from bloat. At least LINUX is free. But all of these share the same modern characteristics of serious hardware requirements (OK, some versions of LINUX aren't so bad here), coupled with high learning curves (yes, even the Mac requires thought now and then) and high support/maintenance/operating costs. OS vendors have assumed that by adding more stuff (the latest being user interface strategies), they become more powerful, but this adds even more cost, as we've seen with Windows over the years.

I have in the past argued that the client OS needs to be doomed because of those very cost issues. Upgrading to Vista makes no sense; it's slow, requires lots of hardware, and has a steep, expensive, and ultimately pointless learning curve associated. Windows Mobile (which I am - gulp - considering purchasing again) also requires significant training or hands-on experience before one can be productive. The original purpose of an operating system was to abstract hardware in order to make it more reasonable for programmers, and also to provide a home for common functions that most programs need. Recent years have seen the rather ugly trend of the OS as strategic competitive weapon, an end in itself, designed to lock out competition, but also lock customers in to a particular platform and user interface, the former an overt trap and the latter simply irritating, especially in the case of Windows Vista. Bottom line, though - the OS might make things easier for programmers, but is it really helping users?

That's why the InfoWorld piece is so interesting. The rise of browser-based apps - which I have long advocated as the right paradigm for mobility - can practically obviate a need for the OS. If the browser runs, the app runs, and OS has no control over the situation; it simply exists to support the browser. All of this is especially important to mobile users because, while we seek to be able to everything we do in the office while out and about, we don't really want to carry a big PC with us everywhere we go. Even MIDs can be too big if you have to lug them around all day. Indeed, what we really want to carry - or perhaps even borrow - is an interface to our personal and enterprise IT resources, some accessed via wireless and some via a USB key. That may indeed involve an OS, but we shouldn't need to carry one everywhere we go - especially Microsoft's bloated, selfish vision of what an OS is or should be. Just for grins, though, consider what may become a very important trend during 2009 - the network or vitual OS. Have a look at G.HO.ST, CloudWorks, Glide, and gOS, just for starters. If this approach catches on, and I think it will over time, the nature of mobile devices changes markedly - and, in the bargain, much for the better.

I'm sure we'll continue to argue this - and many other elements of wireless and mobility - well into 2009. Which, by the way, is here. A happy, safe, and prosperous New Year to all! Yes, you too, Microsoft.

• Wireless / Mobile
• cloud computing
• Linux
• Mac OS X
• operating systems
• virtual OS
• Windows