I've thought for a long time that Red Hat is playing this just right. Anyone in a competitive
undertaking has to work from their strengths--build on them--if they are going to have any
chance of winning. If I were in the Linux distribution business, I do would do the following:
1. Start with servers. That's where Unix has already been, and people are willing to pay for
reliability and support.
2. Become viable as a workstation OS. (CAD, EDA, and desktops for programmers.) This has the
same upsides as #1, and it's good desktop practice.
3. Become the OS of choice for embedded work. This pays, but not as well. It also gives
people another reason to use your OS on a workstation.
4. Target large, restricted deployments, like phone centers and kiosks. These also pay well,
and it's more desktop practice.
5. The enterprise and consumer desktop. I don't think you can maintain 1-4 without a presence
I'm just a programmer; I could have it all wrong. (I also fudged the difference between "OS"
and "distribution", but that's inherent with Linux.) Red Hat, however, seems to be doing
exactly what I would be doing. And I would say that they're between 3 and 4. Let's be
patient with them, shall we?