The thing that always bothered me about de Icaza is that he should know better, given his credentials and how immersed he'd been in the FOSS scene for so long.
He boils down the issues with Linux on the desktop to what jwz calls Cascade of Attention-Deficit Teenagers[0] model. Essentially open source developers don't have the discipline and patience to do the hard work of maintaining their software, and instead just want to refine their designs, throw things away, start over from scratch, and make things perfect.
I don't disagree with this view, and I think it's just somewhat silly to expect the same level of polish and back-compat on Linux that you'd see on a commercial OS with commercial apps. Certainly there is a lot of very polished software built under the OSS model (though many of that software has funded full-time developers working on it), but any project that is either run largely by volunteers, or largely by the programmers themselves, is often not going to end up like a polished, seamless, corporate product. Companies (and product managers) make decisions about building software in a very different way than developers do. They have different priorities, and different things they care about.
There will never be a "year of Linux on the desktop", because FOSS Linux-based OSes are constantly-moving targets run by people who generally aren't getting paid to do the work necessary for that to happen. That's pretty much always been the case, and even with all the corporate interest around Linux (Canonical comes to mind), it's just not happening.
My first experience with Linux was with Red Hat 4 back in 1997 or so, though I didn't start using Linux as my daily driver until 2002 or so. It has always had, and will always have, many rough edges. For me, I find that I have fewer problems with it than I did during my stints running macOS, but... that's just me, and I can fix nearly any issue I run into (even if they are few and far between) with a minimum of effort and time. That's not for everyone, and that's fine.
[0] Don't click (copy/paste into a new tab), as jwz has a nasty redirect for people coming from HN, but: https://www.jwz.org/doc/cadt.html
A lot has changed, Apple braking compatibility, Microsoft braking so much. Microsoft hegemony on desktop destroyed by web, smartphones and Apple, so much that ChromeOS exists. No Flash, no IE, Microsoft Edge based on Chromium. Windows adopted Linux with WSL. Wine getting better, Valve Proton drives gaming on Linux. Open source AMD GPU driver. Wayland. The future is awesome.
He thought of stomping alternatives, parroting Jobs, that's wrong. The reason I've switched to Linux is such attitude from Microsoft. Linux experience is a moving target. At first we are expats, striving to replicate what was lost, but Linux provides much more. Why not explore it?
Microsoft Windows is powered by legacy, enterprise and gaming. Apple macOS advertises polished experience, creative applications, iOS development. Google Android, Apple iOS — touch oriented OS, app store. Google Chromebook — security and web. Linux is different, every community strives to find its own answer.
I estimate that by the middle of this decade, linux will no longer be a moving target for third parties. It is just getting complex, mature and polished enough to keep continuously changing. It will likely be good enough for most things most people would like to do.
Nevertheless, I don't think that will make its market in increase significantly. There are more factors that influence the success on the desktop beyond polish, maturity and technical excellence. Even listening what industry say they need is a good indicator of what needs to be done. The "third party software industry" is like steve jobs said about users: "we can't just ask for what customers want".
