A vendor can often be their own worst competition if they create good technology, but put it out in a way that is too limiting, in platform support or licensing, than their prospective users would like it to be. I've often refered to this as the Anti-Market among colleagues. The rules of the Anti-Market are more or less as follows:
If you create a technology that is useful, but 90% of your prospective market can't use it for various reasons, they've got a good chance of getting together and writing a replacement for your product.
Example 1: KDE vs. Gnome
Gnome created out the anti market that KDE created. KDE is built on QT. Back in the early days of KDE, QT was licenced in rather funny ways by Trolltech. The funny license meant that Red Hat (and other Linux distros) didn't want to ship it. Mandrake was originally just Red Hat + KDE to fill such a need. But with the bulk of the KDE user market blocked because of bad licencing, a void existed to be filled. Gnome did that. A decade later Gnome is the primary desktop environment on nearly ever major distro, and while KDE 4 has gotten some recent press, it is definitely now a minority player.
KDE was brought down because it created an anti market. People wanted that kind of function, but the way it was delivered was not acceptable to its users.
Example 2: Java vs. Mono on the Linux Desktop
How many Linux desktop apps are you running right now, or ever, that are Java based? How many that are Mono based? The only Java apps I run on the desktop in any frequency are Azureus and Freemind. On the Mono side F-Spot and Tomboy have seen a lot more use. Until very recently Java remained under a license that made including it with the Linux platform quite an issue. Mono is under an MIT license, and has been since day one. While Mono has a number of short comings, the fact that it's so young, and so much more used than Java in the Linux desktop space speaks a bit to the anti-market that Sun created by waiting forever to open source their baby.
Example 3: MySQL vs. everyone else
In 1995 Linux was already being used to run key parts of the internet. None of the traditional ISVs were paying attention to it (DB2 showed up in 1998 on Linux, and too my knowledge, was the first big database vendor there). You know what you need to run the internet, a reasonable database. MySQL popped out of the anti-market created by there being a platform people were using quite a bit, but lacking ISV support. People needed the function, but couldn't get it even if they wanted to pay for it.
I continue to be amazed at how much of an anti-market MySQL took advantage of.
The Linux Desktop space is full of anti-market applications, some of which have even seeped back into the Windows world, like OpenOffice, Gimp, and Pidgin. Adobe just made a very astute move and got Air out for Linux before they forced a new anti-market there. While the Linux Desktop space isn't the highest volume space for users, the developer to user ratio in the space is very high, which means ignoring it means there is a real chance of creating an anti-market.
I'd love to hear other people's thoughts or examples here, comments are open, have at it.