Re: A different definition of MINUS, Part 1

Brian Selzer wrote:
Just some food for thought:

1. The scope that McGoveran referred to is the set of all tuples that appear in any possible value for the relvar.
...

I's rather say he was talking about the scope of a complement, a standpoint that seems more useful to me.

2. When dealing with an update, there is always a before and an after, unless the update is null. If "Paul is asleep," and the update states, "Paul is awake," then which is the case? Is he asleep or is he awake? He can't be both. An expression of the algebra can only deal with one set of circumstances--not both, since the algebra is for determining what is the case rather than asserting what is the case. Each possible value for a database corresponds to a distinct logical proposition, yet only one of those propositions represents what /is/ the case. So whenever there is an update (unless it is null, of course), you are always dealing with two distinct propositions representing /what has been the case/ and /what is now the case/. I just think that ignoring before and after invites dire consequences.


You are really talking about language implementation and its various psychological aspects, introducing all kinds of notions I don't think are important to the logical question of equations that represent the differences between certain relations. I'm only talking about the A-algebra's logical implications in the RM. You may want me to talk about language but I won't, at least, not much.

.