Closed source and reinventing the wheel

It's not great news to anyone that closed source code leads to permanent reinventing of the wheel. Nonetheless, I am impressed by the number of programs out there, especially on the Mac platform, that all do the same thing in only slightly different ways. There are at least a half dozen each of RSS readers, notebooks (think OneNote), launchers and web browsers for Aqua/Cocoa. I've no doubt there is a lot of innovation going on and that there are advantages to the approach in that it enforces having completely separate projects, each with complete freedom to develop (after all, some basic components, such as html rendering, are usually provided by the OS, meaning developers need merely wrap a GUI around it). However, it does strike me that the creators are often unaware of other people's existing creations. This does not usually happen with open source, as first versions are released early, and newcomers tend to join the fastest-developing projects, so that even competing lead developers sometimes abandon their projects and join their prior rivals. Nonetheless, as we see with jukebox applications, projects periodically replace each other as something better gets developed from scratch (for an example, see how Amarok is taking over from XMMS - with several others waiting in the wing for their chance - or how gnome is replacing KDE as the main Linux distributors' favourite). I have yet to see this approach stifle innovation, although it has to be said that OS X remains the most technologically advanced environment.