It's funny that we've come so far in computing and yet reliable data storage is still such an unsolved problem for 99.9% of the world's consumers. All the majo

It's funny that we've come so far in computing and yet reliable data storage is still such an unsolved problem for 99.9% of the world's consumers. All the major solutions have their drawbacks. For example, btrfs' present day raid code is in such bad shape that apparently it needs a total rewrite. zfs' volumes can't grow without a resliver, which is not something a home user should ever have to do. From a design perspective, the internals of both btrfs and zfs are only accessible to experts and totally opaque to their users. I've long thought someone could come along and write a FUSE-based distributed storage layer that sat on top of existing file system abstractions like ext4, etc. And that a system built that way could just crush both of those solutions in almost every conceivable way. I've toyed with writing it myself. This approach might not be the fastest, but it would be the most flexible, possibly very reliable, and as a bonus it would be really easy for average users to grok and maximize their data recovery from all but the most catastrophic events.