Interesting read, and a pretty accurate summation. The current lookin'-for-a-patent-infringement-lawsuit climate doesn't do much to encourage innovation, either.
From what I've seen though, it's pretty much the same kind of progression that's happened with so many other areas of innovation. There's always a lot of diversity in the beginning as people approach a new technology from different angles, resulting in a lot of variety, but usually at the cost of overall reliability. Eventually things clump into fewer and fewer schools of thought, as early innovators are fall out of the picture due to lack of business skill, poor product, or any number of reasons.
Just as people decry the lack of variety in automobile design, music, television programming, etc... we see the same basic chain of events in the area of computer operating systems. It's not so much that there's a lack of fresh ideas, they're just showing up in other areas. As the article pointed out, OSes have become increasingly complex, and people and businesses (where most of the computing dollars are spent) don't have the patience or time to go with a rough-around-the-edges beta product for running their business or family finances. The guy who might have worked on a new OS 15 years ago now writes code for applications to run on an established OS. You don't have to look any further than this site to see the proof of that. Object Desktop components add the end functionality and customization that is lacking in the basic OS. Of course we'll see more of those features built in to newer versions of said OS (MS Windows in this case).
Mac was the first to go with a more user-friendly GUI. Microsoft developed Windows to give the IBMs and clones that Mac usability, built on top of DOS, and eventually evolving into the NT platform that 2000 and XP are built on.
Linux still holds a lot of that early development appeal for many, but they still strive for the same basic windows-type environment. The average PC user doesn't want a lot of the extra work that goes into setting up a Linux desktop, and they want to run all their familiar windows compatible applications. Linux has the geek-appeal factor, Mac has the artist/writer, somewhat elitist appeal, and MS Windows, for better or worse, has become the de-facto standard that people expect in their home PC.
For all the innovation MS has brought to the desktop, it can be argued that they have squashed even more by their business tactics. Who has the time, funding, or desire to go up against such a behemoth? Mac seems content to cater to their niche share of the market. Linux, while holding a lead in the server market, lacks the marketing infrastructure and polish to make them as ubiquitous as MS Windows.
That leaves everyone else to fall in the cracks as an also-ran, their innovations either ignored, or swallowed by one of the bigger fish.
