Why Miguel de Icaza is Wrong About Desktop Linux

My take: "The Desktop will die WAY before Linux dies". Nuff said.

>This is where I think Miguel's got the most traction, it's a historically weak area for Linux. But it's quickly becoming less of a problem. "Cloud" apps remove the issue of client compatibly breakage, Google's very successful internal use of Goobuntu proves this, and is an indicator of future business application.

At best that argument translates to "compatibility is not such a problem anymore with cloud apps". Maybe so, but that's hardly an argument for "desktop linux" not being dead. Thin clients and the "cloud" mean that it hardly matters if you are using Linux or not.

And it's not that clear cut that the cloud is wining either. For some stuff it is, for other stuff native trumps it, and he have seen a heavy switch back to native, even for mobile apps. The web and HTML being what it is, there are tons of apps that won't feel any good in cloud form, and those apps are what the desktop is all about.

>Native client apps can be distributed through Ubuntu's Software Center, and they can target LTS releases of Ubuntu for guaranteed compatibility.

It doesn't matter, since that doesn't mean you get the commercial apps that OS X/Windows have. Even the LTS releases provide a too small period of stability for a platform with such small market share.

>He fails to notice that innovation in OS X has begun to slow, and age. It's HFS+ file system is disastrously out of date

Which doesn't matter at all. For the things people use desktop computers for, HFS+ is perfectly fine. Journaling, extended attributes, ACL, Unicode, are all there.

Just because some geeks drooled over the possibility of ZFS on their desktop doesn't mean it solves any real problem desktop users have (oh, and if some desktop OS X user objects to this, let him check the terms "outlier" and "statistical noise").

Plus, even if it was true that "innovation in OS X has begun to slow", it doesn't matter, because it already has a huge, consistent, feature complete and widely deployed code base out there already. And, as Miguel said, too much innovation is one of the problems with desktop Linux.

>and OS X remains tied to hardware that's priced beyond what a large market will pay.

Actually OS X sells better than ever every year, and it got from 2% to 10-15% of the market. Linux being available on cheaper hardware (and free) itself, didn't help it. So the argument is moot, except in developing countries and such (where most of the population will rather pirate Windows anyway).

>The future of Linux on the desktop, in the context of something that's targeted by commercial parties and has years of stability expected, already has become Ubuntu LTS.

That's a "too little, too late" kind of stability. The LTS kind of releases are of more interest to server deployments and the enterprise, anyway, than in any commercial party interested in desktop apps. And we were speaking of "Linux on the desktop here".

E.g Adobe won't release Premiere for LTS only to see that in the next LTS base Linux/Gnome APIs and frameworks have been rewritten and replaced with something else "just because".