cat /dev/DiBona/brain: Considering Mono
This may be stating the obvious for a number of users, but I thought I'd spend some time pointing out that Microsoft needs Mono a lot more than it likely is comfortable acknowledging. We Linux users can become very single OS-focused at times, thinking of the world as being black (XP) or white (the penguin, natch). I do not mean this as a criticism per se, but it is something I myself have been guilty of, having barely gotten my toe wet in the FreeBSD waters. So, I thought I'd project it onto a number of my colleagues in the Linux community--you don't mind, do you?
Languages, however, don't have this luxury; they must work on as many of the major platforms as they can to attain the popularity and ubiquity necessary to be used on a large scale by professional software developers. Exceptions abound, of course. Consider, for instance, embedded languages--small c and the transputer language Occam (specifically designed for old school multiple processor inmos transputer cores). But, for the most part if you look at how languages spread out and become popular, you can see them spread from platforms ranging from the largest machines to the smallest handhelds.
If your goal is ubiquity--and when has ubiquity not been Microsoft's goal?--then the ability to run your code on other platforms is vital. Thus, Mono is vital if Microsoft wants .Net to overtake Java, which it seems to want real bad, but perhaps not bad enough to make a version for Linux, BSD and the other major UNIX platforms. For Microsoft, reaching out to UNIX seems to include only investing in it and not actually porting tools.
My excuse: I've been thinking a lot more about .Net and Mono lately, because I am advising a company using Mono as a backend framework for a really cool piece of software. I've also looked into the suitability of the language for a teeny side project that I'm looking at as an excuse to learn the language.
After some consideration and a little bit of coding, I've found there are some very real barriers to cross-platform .Net adoption. I'm thinking mostly of the lack of a cross-platform UI toolkit. The only people addressing this lack seem to be the Mono guys. Thus, if .Net is to succeed as a preferred language where Java currently holds sway, then Mono is a vital part of that equation. If it didn't already exist, Microsoft would have to make sure that it did.
Regarding user interfaces, GTK for Mono isn't there for all platforms and winforms running on top of winelib, and the rest is underway. Is that an installable, maintainable way of providing winforms? It strikes me that this might be too brittle a solution, but I haven't found any real holes in the approach. If you have, say so in the comments, but this could be a real barrier for software with real user interfaces attached. A Mono winforms-compatible layer is key to Mono being able to compete for desktop applications and for Linux to act as a first-class platform for .Net applications.
The question for the Free Software community to ruminate is this: Is supporting Mono supporting our own desktop downfall? A statement attributed to Steve Ballmer at a party a few years ago would support this possibility. To paraphrase, when asked about Mono, he shrugged it off with a statement alluding to the fact if Mono got too popular, that's what lawyers are for. Mind you this is friend-of-a-friend kind of hearsay, but it is worth noting regardless.
Because Mono works off the ECMA spec for C# and the CLI, and because Novell has the legal resources to make the idle-lawyer threat less effective, I don't think this is some long-plotted submarine trick designed to disrupt free software adoption. In the long run a healthy Mono that can run .Net apps natively is as important to Linux as a good JVM is--that is to say, very important.