Temporal generalization comes in three guises. Consider a temporal database in which data items are timestamped with valid and transaction time. Temporal generalization occurs when weakening constraints hitherto applied to the timestamps. Used in this sense, temporal generalization is the opposite of temporal specialization.
Next, a temporal relation is generalized when new timestamps are being associated with its tuples. In larger information systems where data items flow between multiple temporal relations, items may accumulate timestamps by keeping their previous timestamps and gaining new timestamps as they are entered into new temporal relations. Thus, a tuple in a particular relation has multiple timestamps: a valid timestamp; a primary transaction timestamp, which records when the tuple was stored in this relation; one or more inheritedtransaction timestamps that record when the tuple was stored in previous relations; and one or more additional timestamps that...