It might also help to briefly consider why this concept of a division of space exists in the first place ... Let me paint a simplified picture.
In a very large, production database, it might not make sense to put all of the tables in a database on a single disk drive. Doing so would cause fairly murderous amounts of movement of that hard-working read/write head as it scuttles back and forth across the disk surface, slowly wearing out its bearings while performing only read/write operation at a time. ZZZzzzzz....
Instead, tables are spread out onto a "farm" of separate disk drives. Now, it becomes possible to efficiently do more than one disk I/O operation simultaneously, and the read/write mechanism on any given drive moves much less. The speed of the database is greatly increased, and the wear-and-tear on the drive farm is greatly reduced.
On a simple, IDE-style disk subsystem, not much advantage might be perceived, because IDE can't do much in the way of simultaneous disk I/O. But subsystems like SCSI and fiber-optic channels can have dozens of operations going all at once. This is where tablespaces become very important.
Segments collectively form a Tablespace. There are different types of segments as given below:
Data Segment: Stores data information
Index Segment: Stores index information
Rollback Segments: Stores undo information.
Temporary Segments: Uses this segment when ORDER BY or SORT or any other operation is being used by SQL statement.
Tables form a part of Tablespace. Tables can be spanned across the drives.
Hope this much helps...