Complements
Date: Sun, 21 Jan 2007 17:05:31 GMT
Message-ID: <v3Nsh.748776$5R2.393636_at_pd7urf3no>
Another thing that has eluded me for some time is something David McGoveran said on the dbdebunk site near the bottom of:
http://www.dbdebunk.com/page/page/1396086.htm
(quote)
The "problem" of finding a unique inverse transformation for a join view
in a given context without specifying any context goes away. Clearly,
the delete must satisfy 'NOT PA OR NOT PB'. (Aside: We must be very
careful what we mean by NOT, avoiding confusion with simple complement.
Relations have a relative complement - tuples not asserted to be 'true'
- with a scope that is distinct from that of the relvar predicate -
tuples that cannot belong to the relation.) This requirement may be
further constrained in at least one of two ways. First, if they know,
the user should specify their intent as to why the conjunction is
'false'. Second, some of the integrity rules and dependencies the
designer imposes may further constrain the general requirement.
(end quote)
I have presumed that an implementation that supports the closed world assumption could allow tuples in a negation complement that if negated again could never be a member of that complement's relation because of various constraints such as keys.
Does anybody think McGoveran is saying otherwise? Or that he is suggesting something other than this kind of CWA complement?
p Received on Sun Jan 21 2007 - 18:05:31 CET