Linux is a great operating system. In fact, it is as good a UNIX clone as there ever was. Lots of innovation has grown out of the open source movement (ironically born out of AT&T's efforts to further the development of UNIX during the 1970's). Those early efforts to improve upon UNIX led to BSD (and many other flavors of UNIX). In fact, Solaris (formally, AT&T System V, Release 4) was co-developed by AT&T and Sun Microsystems in response to the overwhelming success of all those non-AT&T-developed flavors of UNIX.
All these efforts led to remarkable innovation during those twenty years before Linux hit the streets -- and the 14 years since! The open software movement born on college campuses across America led first to Gnu (Remember their slogan? "Gnu's Not UNIX") and ultimately prompted Linus Torvalds, in the 1990's to develop the Linux kernel to finally free all those Gnu utilities from the encumbrances of UNIX. The rest is history.
From 1991 through 1996, I was a UNIX systems administrator. At home, I piddled extensively with Slackware but I soon grew bored with the experimentation required to use Linux. (In those days, OpenOffice didn't exist so I was still tied to Windows for much of my personal productivity needs.) Nevertheless, I never stopped watching Linux.
After my five-year foray into UNIX/Linux, my job responsibilities changed and I returned to my DOS/Windows roots. I soon realized that Windows' future was in a UNIX-like implementation, including full preemptive multi-tasking with robust task and resource management. Apparently Bill Gates recognized this too (see http://en.wikipedia.org/wiki/Windows_nt). No matter. On to my point ...
Linux fans (and start-ups, like Red Hat) spent most of the 1990's positioning Linux as an alternative to UNIX and that has paid off. Today, there are few applications where Linux cannot do as good a job as UNIX can. The IT community knows this too; UNIX and Linux systems stand side-by-side in the machine rooms of most fortune 500 companies.
Of course, Microsoft spent no time dragging its feet either and by the late 1990's Windows Server could do most of the things (but certainly not all) that UNIX and Linux could do in the machine-room and, because Microsoft was busy developing proprietary services which only Windows Server could deliver to Windows desktops (and protecting those services with patents, not copyrights), Microsoft made it's way into the machine-room right next to those UNIX and Linux servers.
There is no doubt that without the open source movement, the level of innovation we see today would not be present anywhere in IT. If IBM had their way, we'd all still be tied to mainframes and dumb terminals -- and to paying monthly maintenance to IBM to keep them running.
In order to break out of the machine-room and onto the desktop in any meaningful way, Linux vendors have to take the next step. They have to get their OEM partners (that's IBM, HP/Compaq, and Dell at the top -- and a handful of second-tier OEMs) to start marketing Linux to consumers.
Many of you ask why and the answer is simple. Linux has gotten where it has almost entirely through word-of-mouth. In machine-rooms around the world, which are full of people with years of experience with IT, that's not hard to do but in the consumer space, where everyone has their own take on their favorite brand, even the most honest presentation of the choices available leave the untrained full of fear, uncertainty, and doubt.
If Linux (or UNIX, for that matter) is to break out of that "IT inner circle" it must be comprehensible to the average consumer with no IT experience whatsoever. It must also be able to run the most popular applications from Adobe, Intuit, and many others. ISVs will not port code to Linux unless they see market potential in such a move. Without aggressive marketing by OEMs, the market potential of Linux will never be realized.
Ask the developers of your favorite Linux distribution why you can't buy a computer with that distribution pre-installed. You just might be surprised at their answer.