Re: object algebra
Date: 24 Feb 2004 22:01:39 -0800
Message-ID: <4b45d3ad.0402242201.72ab2c26_at_posting.google.com>
> If Mary has no eyes, then the definition of a person as having eyes is
> flawed.
Since at design time, it is not possible to know the future, according to the above, in RDM a relation becomes flawed by the entry of a tuple that don't supply all the values. What if a tuple has an extra/new attribute? (In either case, this flaw doesn't occur with TDM)
> However, if it's convenient to require some designation for eye
> color, NULL is a lousy way to do it. You should in this case have an
> EYE_COLOR type, and one of its values could be "Not Applicable" or "Unknown"
> (both of which have well-defined meanings, unlike NULL).
NULL kills closure. How does substituting well-defined values such as "Not Applicable" or "Unknown, correct the problem?
Also now the domain/type EYE_COLOR has tuples "Not Applicable" or "Unknown". Doesn't this violate basic set theory?
Is "Not Applicable" a color? Received on Wed Feb 25 2004 - 07:01:39 CET