I think a combination of tagging, auto-classification (like in DevonTHINK) and a probability-based scoring system with heavier weights given to time and frequency should do well. i.e. documents that were created/modified/read recently, and those that are frequently accessed. Kind of like email. :) The file system still remains, but ideally no one should be looking at it in its raw form. Managing a hierarchical file structure is a major pain.
I think a combination of tagging, auto-classification (like in DevonTHINK) and a probability-based scoring system with heavier weights given to time and frequency should do well. i.e. documents that were created/modified/read recently, and those that are frequently accessed. Kind of like email. :) The file system still remains, but ideally no one should be looking at it in its raw form. Managing a hierarchical file structure is a major pain.