UDI: Why did it fail

Peter

I know source compatibility is the goal and does work, but this isn't enough given that:

1. Source compilation is a pain for the average user, at least w-o a good front end that resolves dependencies, etc. and 2. Hardware manufacters might not want to release open-source drivers and a binary driver is clearly better than no driver and a good binary driver is IMHO better than an open-source driver that's a reverse-engineering job and therefore probably either full of bugs or missing features, or in the best case takes months after the hardware is released to write.

Also, there's nothing that says UDI had to exist in the main kernel tree. If Linus feels it should be developed outside the main tree, so be it. It could have simply been included in the more desktop-friendly distros (since it is really unnecessary in server-supercomputer distros or in embedded system distros, where in either case users don't have a wide array of off-the-shelf hardware that they expect to "just work"). It could be included in such distros as a custom module, or hardware manufacturers could have packaged it in source form wa nice compilation front end (UDI was under a BSD style license) along with their binary drivers.