There are a couple of points that fall out of this interview. Firstly, de Icaza doesn't have the same god-like power over GNOME that Linus does over the kernel. He's influential, sure. But the real "leader" of GNOME is surely Havoc Pennington.
Secondly, I think that if de Icaza (and by extension, Ximian) do go forward with this, GNOME will fork. There will be a Mono version, and a Bonobo version. This will happen regardless of whether it's a good idea or not, because some people are violently opposed to using technology that is in any way associated with or emanates from Microsoft.
The problem (for the Bonobo-based GNOME people) is that Ximian are really, really good at developing GNOME apps that people actually want. There isn't a mail, contacts and calendar program on Linux (or any other operating system, IMHO) to match Evolution. Only StarOffice and OpenOffice hold a candle to Gnumeric. The Ximian Setup Tools are showing severe signs of having been designed for actual normal people to use. Red Carpet is a dependency-solving, channel-based, distro-independent system updater and software installer that actual normal people have a hope in hell of using.
With Miguel driving Mono-based GNOME development, it's highly unlikely that these applications will continue to be developed on Bonobo come GNOME 3.0. Sure, they're GPLed, so you can do the work yourself, but porting from one component architecture to another is a lot of work, and even assuming you could find enough people to do it, you'd always be some number of versions behind. Of course, Ximian could opt to use Mono to generate Bonobo components, and the Ximian Mono-based applications will still be able to talk to Bonobo applications, but I'm guessing (uneducatedly, admittedly) that Mono-generated Bonobo components are going to be less efficiently than straight Bonobo ones. Performance is going to be a big issue.
The question in my mind is, "does it matter if GNOME is based on Mono?", which in turn really begs the question, "What does Mono bring to the party that only Mono can provide?".
If Mono provides new capabilities that are going to be required to build the next generation of net-enabled, secure desktop applications, then there are two options. Build these capabilities into new versions of Bonobo, or go with Mono. All this is, of course, assuming that we want to build these types of applications on GNOME. You can be sure that Ximian are going to, though.
What I'm most concerned about is the possibility of Ximian becoming the de-facto GNOME standard, with the actual GNOME Foundation being regarded as something of an anachronism while Ximian do all the cool beans.
No one company should be in a position to dictate the GNOME development direction, be it Ximian, Sun or HP. That's what I'm really bothered about, rather than the component architecture.
One of the things that attracted me to GNOME rather than KDE in the first place some two and a half years ago was that the development process was truly Free; it could never be taken hostage by this company or that changing this or that licence agreement. I don't want to see that situation change, even with a company as cool and as seemingly proficient as Ximian.
There's also a very cynical part of my brain that is whispering in my ear that the whole Mono project is the first plank in a strategy by Ximian to take control of the GNOME project. In the future according to Miguel, whoever controls Mono will control GNOME. And that, perhaps, is Ximian's goal.