I wear two hats. I am Linux/Unix user and I am a java programmer.
When I'm at home on my Linux desktop, java is far from my mind. I want to make Gnome (or KDE) the best desktop environment there is. For this I think about seamless multi-language development and powerful common runtime libraries to support *nix applications. I dream of the availability of games for Linux and of standardized multimedia support.
I can see why the CRI appeals.
At work I see another side of *nix. Enterprise development for a while now has been dominated by Java. Even better Java deployed to *nix boxes. Around this has grown astonishingly successful open source projects, such as, Tomcat and JBoss.
To my bosses (and most enterprise) open source has been validated by these projects. Open source has been taken out of the domain of hobbyists and become a deciding factor for which platform and which technology to use.
It is my daytime job that MS is after. I have heard it suggested that MS is using the desktop war as a diversion from their real goal of the cash cow, enterprise development and the technologies that defines Internet development. I don't know if I believe that they are that smart :-) But two different goals are getting confused.
One is having the tools to develop killer apps for the Linux and realize Linux as a consumer product. The other is the technologies we use to move to a more distributed way of computing (web services) and what we usually call server-side development.
It is the later that MS is after. The later is also dominated by *nix servers and Java solutions presently.
Although I imagine Mono is thinking along the lines of integrating the two, eg Visual Studio. Is this that important? IMHO there are other much more important issues for the development of internet applications than having it integrated with Gnome.
Java/*nix is already ahead in the enterprise arena. Why give up the heard start? Wouldn't leveraging the technology that already exists in Java be a wiser option? I think that going for enterprise development through an open source .net is going to be a long hard road. The CLI does offer some great possibilities in terms of having a multi-language and a platform independent layer for *nix development. But it is worth keeping in mind that server side development is already being won. Taking the good ideas of .net and extending Java would be the surest solution.
I would like to see over the next few months, Sun opening up Java more, and the open source community considering what projects need to be launched to not just match .net but to offer something unique and special to the open source model.