This is the single biggest failing of the FOSS ecosystem.
Someone starts a piece of software and gets some of the desired features working. Shortly after that, someone else, either working on the project or using it, decides one of several problems plague the program. Either it's development is too slow, it has crummy architecture, someone else thinks they can do better, philosophically or technically, or they are half-baked programmers who look at existing code, can't figure it out, and decide to start over from scratch. Or maybe the project's lead(s) decide that their way of doing things, technically or philosophically, is the only "right" way, and hit would-be contributors over the head with attitude (I'm looking at for example developers of VLC and cdparanoia, not to mention the issue of Linux kernel schedulers and sound subsystem).
So we end up with multiple half-baked programs all doing sort of the same thing in different ways but none of them doing the whole job. Naturally, when someone sees the situation, the first reaction is "All this mess! I'm going to start a NEW project and do it RIGHT this time!"
If we FOSS users and developers are lucky, eventually there will be a tipping point when a majority gravitate to one project and things get more or less sorted out. If not, well, we can always use ANOTHER, say, media player; some college CS major can tackle it as a senior project, release it, and then forget all about it. If Amarok, Audacious, Beep, BMPx, Banshee, Kaffeine, Miro, Rhythmbox, VLC, Winamp, XMMS, xine and whatever else I'm forgetting don't offer enough choice for you.
Glad to see that yet another category of software is joining the party.