Databases Reference
In-Depth Information
The distinguishing feature of a temporal database is, of course, time
itself. Temporal database research has therefore involved much investigation
into the nature of time itself. Here are some of the issues that have been
explored:
Whether time has a beginning and/or end;
·
Whether time is a continuum or occurs in discrete quanta;
·
How best to characterize the important concept now (often referred
to as the moving point now);
·
and so on. But these issues, interesting though they might be in themselves,
are not especially database issues, and we therefore do not delve into them in
this chapter; instead, we simply make what we hope are reasonable assump-
tions at appropriate places. This approach allows us to concentrate on mat-
ters that are more directly relevant to our overall aim. However, we do note
that portions of the temporal research have led to some interesting generali-
zations, suggesting strongly that ideas developed to support temporal data
could have application in other areas as well. Note: This last point notwith-
standing, we follow convention in referring throughout this chapter to tem-
poral keys, temporal operators, temporal relations, and so forth, even
though the concepts in question are often not limited to temporal data as such.
Caveat lector! We concentrate in what follows on what seem to us the
most interesting and important of the various research ideas (in other words,
the chapter is our attempt to distill out and explain the good parts of that
research, though we do depart from the literature here and there over ques-
tions of nomenclature and other small matters). Be aware, however, that little
if any of the proposed new technology has yet shown up in any commercial
DBMS. Possible reasons for this state of affairs include the following:
It is only recently that disk storage has become cheap enough to
make the storage of large volumes of historical data a practical
proposition. However, data warehouses are now becoming a wide-
spread reality; as a result, users will increasingly find themselves
faced with temporal database problems and will start wanting solu-
tions to those problems.
·
Although most if not all of the features we describe have been imple-
mented in prototype form, their incorporation into existing prod-
uctsespecially SQL products, where SQLs departures from the
·
