Stardate 20031110.1222

(On Screen): Remember Netscape? Used to be a big company, used to be a big success. It used to be a contendah... It was part of the dot-com bubble, of course, but in fact for a while it had a pretty good thing going, and its business model didn't seem to be built out of the kind of smoke and mirrors used for a lot of dot-coms which only lasted until the venture capital had been burned up, usually shortly after the IPO. Netscape was well placed to capitalize on the rise of web browsing, being the source of what was widely considered the world's best browser, and an excellent web server package.

The browser was sold to commercial users, it was explicitly free for academic use, and basically shareware for everyone else. If you paid there was some nominal benefit which I don't even remember; it wasn't what motivated me to pay my fee, even though nothing forced me to.

But the people at Netscape had even bigger plans. Their package was cross-platform. And HTML was supposedly a standard governed by an industry body, but as a practical matter Netscape did whatever it wanted and the industry body had little choice but to rubber stamp most of it afterwards. When the browser used by 80% of users had a feature, it was a standard even if the standard didn't say so. So for a while Netscape was able to add more and more features to HTML, making it more and more powerful, until it went well beyond the original concept of a text formatting language, becoming instead a pretty comprehensive way of describing user interfaces which was entirely platform independent. And it didn't have to be used across the internet; it worked just as well with local HTML files, and it could quite reasonably use HTTP to talk to a local program or a program on a department server instead of a remote web server. As such, the local program could do whatever job it needed to do, and could rely on Navigator to take care of all the issues relating to presenting the user interface on whatever operating system it happened to be running on. That meant that it was easier to make those application programs cross-platform, since they wouldn't have to implement an OS-dependent user interface. They'd still have to interface to the OS execution environment and its file system, but GUI porting was always a much bigger issue because the basic GUI concepts varied so radically from OS to OS.

Even more interesting was that Netscape had created a sort of mini-OS-execution environment inside Navigator itself, where "applets" could run. In principle, a given applet could run in Navigator on a given CPU no matter what OS was installed, because the applet only saw Navigator. If that had been expanded, Navigator would become a sort of meta-OS which ran on top of some other OS, and took over the screen and keyboard of the system most of the time. And it would present the same use experience and app-execution environment no matter what OS was installed.

Microsoft dominated the desktop operating system world. From a technical standpoint, its operating systems of the era truly stunk; they were flakey, unreliable, hostile to non-technically-sophisticated users, and offered developers a monumentally bad execution environment for their programs, with a really arcane programming model. (This being the days of Windows 3.1.) But Microsoft also had the bulk of the market, in part because PC hardware was a commodity, and because Microsoft offered its OS for preload to PC manufacturers at very attractive rates (which is part of why PC hardware became a commodity). It's has also been alleged that Microsoft used unsavory and possibly even illegal tactics. However, its biggest advantage in maintaining its lead was the amount of software which was available to run on it much of which didn't run on anything else. Many customers chose it because they needed apps which didn't run on any other platform. Others began to use it because it came preloaded but eventually came to depend on Windows-exclusive apps and got locked in.

That huge Windows-exclusive base of application software was something Microsoft had worked hard to support and encourage, and continues to work hard to support and encourage. Since it is a result of network effect, it tends to be self-sustaining, but Microsoft doesn't leave anything to chance and invests a hell of a lot of money supporting the Microsoft Developer's Network (MSDN), which isn't remotely paid for by membership fees. Those fees are nominal and are mainly intended to filter out riffraff, so they're high enough so that college students won't join but low enough to be considered negligible by legitimate developers. (And in any case, much of MSDN's material can be downloaded even by said college students even if they're not members.)

For a large percentage of customers that app-base was the single most important factor in their decision, vastly more important than the technical merits of Windows as an operating system (compared, for instance, to IBM's OS/2). For anyone to upset Microsoft and to dethrone Windows as the most widely used OS in the world, it would have to run as much software as Windows did, and be better in other ways so as to present a compelling case to switch in the face of natural inertia. But there was a bootstrap problem (a "chicken-and-egg" problem) for anyone trying to do so: app developers wouldn't want to invest in developing apps for an