I guess what I'm asking is: if other OSes/apps are going to be successful, they end up having to march to the beat of the system.... (It's not that much different than the indie v. chain bookstore discussion we just had.
===========
This is a controversial topic in the Linux/BSD world (much more so in the Linux world which contains many people who long to see Microsoft de-throned from its iron rule of the desktop - FreeBSD folks scarcely have this fixation).
The core of the controversy is the question of how much Windows-style functionality should be duplicated - if any - on Linux platforms.
As in other areas of life, there are purists.
These high priests insist users must become accustomed to the challenges of a hardcore OS or remain in the foobar world of Windows - but safely contained somehow, like a crazy relative in the dusty, New England attic of a John Updike novel. Others, probably the majority of Linux enthusiasts, insist that desktop competition should be based upon giving users at least as much as they're now accustomed to in Windows only delivered, under the hood, with greater technical excellence and security.
Smooth usability is the heart of the thing.
At home, I use several machines running, individually (as opposed to dual boots or virtual machines), WinXP Pro, Red Hat 9, Knoppix 3.2, Knoppix STD, Fedora Core 2 and Windows Server 2003. I retain the Windows machines to stay in the Win32 client-server, active directory theory and practice loop - to keep my knowledge of this platform current. My actual working machines are Linux.
I must admit that although Gnome and KDE (particularly KDE in my experience) do a very good job of making the overall Linux experience rather crisp there are quirks (such as inconsistent auto sound card detection - a known problem with Gnome on Red Hat 9) which many non-technical users may find hair pullingly annoying.
There's an abundance of info available to solve these problems - some providing GUI based methods, others taking you to a term session for fun with config files - but it's not really reasonable to expect someone who just wants to use his spreadsheet app to perform this sort of triage.
Recently, after a buggy package upgrade, my X-server took an action packed trip to broken dependency hell. I spent many sub-optimized via red wine minutes remaking and rebuilding from the XFree86 sources. As I was busy typing, sipping and imagining what Julie Delpy was doing at the moment I paused to imagine what my mom - a happy Windows user (since I keep the OS healthy through remote vigilance) - would make of this situation - if, that is, she didn't have a helpful son in the biz.
How would she get her X-server - which she didn't even know existed till it broke - back in a useful state without re-installing Fedora?
When this questioned can be answered in a way that minimizes pain for the non-technical Linux will have made tremendous strides towards more universal acceptance as a desktop tool for the masses.
Clever minds are at work on this problem, even as we speak. Behold, Grokdoc!
http://www.grokdoc.net/index.php/Main_Page
.d.