To expand on what Jim said: the hierarchical structure (i'd rather call it recursive) comes from the fact that (if we take a somewhat simplified outlook) UNIX filesystems consist of two things:

- files
- directories

where the directories themselves consist of other directories and files (this is where the recursion gets in).

There are other (non-hierarchical) concepts to handle disk space:

in VM/CMS and its descendants (IBM mainframe computers) there are single disks (informally called "3380" or "3390" devices because they are virtual disks coming from such devices) without any directories. You "mount" (=attach to a virtual machine) such a virtual disk with a certain letter and this letter becomes the last part of the filename, which consists of 8 characters plus 8 characters extension, i.e. "FILENAME TEXT A2" which is a file named "FILENAME" and with extension "TEXT" residing on a disk mounted right now as "A". The "6" is the filemode, which tells the system how to handle a file (for instance there is one meaning "remove after being read", which is used for incoming mail).

In Windows and its predecessor DOS there is a somewhat hierarchical system in place but there are "drive letters" breaking up this hierarchy. While a UNIX system will always have exactly one tree-like structure, regardless of how many disks are used to create that filesystem in DOS there will be a letter for each disk (non-hierarchical) but a hierarchical directory-structure within it.

In OS/400 (another classic IBM midrange architecture) there is no filesystem (in the classic sense) at all, because the system uses its built-in database (DB/2) to store and retrieve files. Basically one retrieves information using some sort of SQL. Furthermore the system makes no difference between memory and disk: everything is one flat space and to write a file to disk ones moves it from one location (which represents memory) to another (which represents a disk or maybe some other device).

I hope this helps.

bakunin