What users will see in the next few years is a replacement of the filesystem storage paradigm with a database storage paradigm. Microsoft is working on it and Apple seems to develop something similar as well. Oracle already stretched their database to accommodate some kind of filesystem like access – which is not aimed at turning a database into a filesystem but rather to offer a transition path.
Replacing the filesystem with a database will allow content to be managed based on wider attribute sets and mostly independently from the file format. It will be possible to look up content on your drive similar to SQL queries. Finding files and/or content on hard disks will be much faster then today – and it may require a lot of more capacity per content for the metadata.
The new Windows version codenamed “Longhorn” is supposed to be finished two years from now – but we can already see that the database paradigm fundamentally changes the basic principles of standard user interfaces: the on screen items are less about managing files but rather about managing tasks. Files (or objects) will always be related to other files, contacts, tasks, dates, places, and so on. The graphical user interface will become object oriented.
If you knew me for longer you’d know that I have always been a supporter of the idea to hypertextualize the desktop and offer users more ways to describe context and relations. The database paradigm is a technological prerequisite to achieve this and to finally realize Ted Nelsons familiar vision of a hypertext OS.