A Linux users' reflections on the Linux part:
You start with:
However, one area where Linux will never succeed is the desktop.
This can be set against a key sentence on the next page:
The only way Linux can reach the mainstream desktop is if that uncertainty is removed.
Based on this and the article as a whole success is equivalent of mainstream. I think that's an example of flawed logic. In many other technical areas we're used to have different niches, but when it comes to computer operating systems for desktops it suddenly becomes a question of being mainstream, not about technical strength or the diversity of user preferences.
It's kind of strange that a user of Linux, I'm refering to the author, falls into the trap of seeing only two extremes: "technology geeks" and "mainstream users". The interesting conclusion of this is that the world of computer users already consists of a strangely large portion of "technology geeks". According to gathered statistics it seems like Ubuntu and Fedora alone have about 18 million users at the moment. Don't try to convince me that all of those are "technology geeks".
I don't expect any special Linux year on the desktop, and to me that's an irrelevant question. Even if we only limit our interest to actual statistics, should every product, in this case operating systems, on the market attract mainstream users? To me that sounds like a malfunctioning and boring market. Let's say only 10 or even 5 % will be attracted by Linux for desktop, why is that a failure? Or why should Linux try to be just another OSX or Windows?
That Linux is little bit of everything is true, but it's also fair to add that the user is enabled to perfect it to his liking. Hence I agree with the conclusion of what possible can be done, but not under the already flawed pretext of "the only way Linux can reach the mainstream desktop".
All your suggestions are good in a way, but there's no need to become overly concerned about being mainstream. Why? To really see the beauty of Linux demonstrate on a bad ass system and show how it makes those CPU and GPU cycles kicking ass. Attract more of those not particularly geeky users, but folks with a profound interest in computing. How will that benefit Linux? Because every user of Linux has a tendency to interest somebody else, and that single Linux users with fairly good talent for computing will be far better than any support you can get for Windows. Windows support isn't very good for the mainstream market, if it doesn't cost you money you run the risk of getting a lot of bad advices.
Maybe it will eventually become easier to make Linux succeed on the desktop as a business model. First though lets stimulate the community interest, let the army of Russians who will grow up using Linux at school dig in deep and contribute, because with a little bit bigger young user base there's nothing to worry about. Linux isn't dependent on 5, 10 or 50 % of users because on average a single Linux user contribute far more than even groups of Windows users.
Mainstream isn't the only option.