I’ve observed, in broad strokes, a pattern over the past 20 years or so.

At one time there was DOS (though I was a happy VAX/VMS developer at the time, and so didn’t suffer the pain of DOS).

Then there was Windows, which ran on DOS (sadly I did have to deal with this world).

Then we booted into Windows, which had a “DOS command prompt” window, so DOS ran on Windows (and the world got better).

Windows lasted a long time, but then came .NET, which ran on Windows.

I always expected that we’d see the time come when we’d boot into .NET, and Windows would run on .NET.

I no longer think that’s likely. Which is too bad, because that would have been cool.

But in the late 1990’s there was the very real possibility that the browser would become the OS. The dot-bomb crash eliminated that debate, robbing the browser-based company’s of their funding and allowing Windows, Linux and Mac OS to continue to dominate.

In the past 2-3 years though, we’ve seen something different. Flash/Flex/Air, Silverlight and Google Gears have shown that the “browser as the OS” idea didn’t go away, it just went to sleep for a while.

Now I’ll suggest that there are really two camps here. There’s the Google camp that really wants the browser as the OS. And there’s the Adobe/Microsoft camp (can I put them together?) that realizes that the technology of the web is really a stack of hacks, and that it is high time to move on to the next big thing. My bias is clearly toward the next big thing, and has been since the web started being misused as an app platform in the 1990’s. They are proposing the mini-runtime-as-the-OS concept, using the browser as a simple launch-pad for the real runtime.

But either way, the concept of a “client OS” has been fading for some years now in my view. Or conversely, the concept of “all my apps run in ‘the browser’” has been ascending steadily and rapidly over the past 2-3 years.

And I’m OK with this, because we’re not talking about reducing all apps to stupid browser-based mainframe-with-color architectures like the web has tried to do for over a decade.

No! We’re talking about full-blown smart-client technologies like Silverlight, where the client application can be an edge application in a service-oriented system, or the client tier of an n-tier application. We’re talking about the resurgence of the incredible productivity and scalability offered by the n-tier technologies of 1997-98. We’re talking about getting our industry back on track, so we can look back at 1999-2009 as nothing more than a “lost decade”.

(can you tell I think the terminal-based browser model is really lame?)

Does this mean that Windows, Mac OS and Linux become less relevant? I think so. Does this mean that Silverlight is the most important technology Microsoft is currently building? I think so.

Consider that Silverlight, in a <5 meg runtime, provides everything you need to build most types of powerful smart-client business applications. And consider that these applications auto-deploy and auto-update in a way that the user never sees the deploy/update process. Totally smooth and transparent.

Consider that you can write your C# or VB code and run it on the client. Or on the server. Or both (see CSLA .NET for Silverlight if you want to know more). You don’t need to learn JavaScript. You don’t need a totally different language and runtime on the client from the server. Silverlight provides a decent level of consistency between client and server, so you can reuse code, and absolutely reuse skills, knowledge and tools.

Google is totally on the bandwagon of course. Many people have seen their strategy for months, if not years. And we’re now seeing it come through, as they start to make their browser is the OS play in earnest.

What’s interesting to me, is that in either case – the browser-as-OS (Google) or the mini-runtime-as-OS (Silverlight) – leaves the previous generation OS concept (Windows/Mac/Linux) out of the loop.

Windows 7, perhaps the best version of Windows ever created, might be the last great hurrah for the legacy OS concept.

I mean all we need now is a Silverlight-based version of Office (to compete with Google Apps) and the underlying “OS” beneath Silverlight becomes rapidly irrelevant...