A bit of a wish list, I know, but does anyone know of a tool, that can anlyse an NTFS filesystem and create an instance of any duplicate file in a sperate folder to link the other instances to using ln. For example, consider the directory listings: D:\notebook01\c_drive_backup\file01.doc D:\notebook01\c_drive_backup\file02.doc D:\notebook01\c_drive_backup\file13.doc D:\notebook01\c_drive_backup\file16.doc and: D:\notebook02\c_drive_backup\file03.doc D:\notebook02\c_drive_backup\file04.doc D:\notebook02\c_drive_backup\file13.doc D:\notebook02\c_drive_backup\file19.doc If it were found that the file common to both lists (file13.doc) had the same size and md5 hash we could create a folder with that hash name, copy the file into it and link the originals to it with ln for single instance storage. The other issue is to mop up 'orphans' that are no longer referenced elsewhere in the filesystem. Obviously there exists filesystems like Sun Microsystems ZFS which can take care of this stuff for you on the block level, and indeed, you can run NTFS over the top of ZFS, but the point here is to do it cheap. Any advice appreciated. Denys Williams
