The solution to solving/maintaining/evolving the schema problem is, to
not have a "schema" in the first place as in xrdb. Instead, let the
current "schema" be defined by the current data in the db. If a db has
ten persons and each person has wildly different attributes, what is
the schema? And what will be the schema if each next person is again
different than any of the prior? Having and/or expected things to be
the same is a special case, it is not the general case.
Received on Mon May 30 2005 - 22:02:32 CDT
