# Re: more on delete from join

From: Marshall <marshall.spight_at_gmail.com>
Date: Mon, 31 Aug 2009 07:49:00 -0700 (PDT)

On Aug 30, 11:52 pm, "Mr. Scott" <do_not_re..._at_noone.com> wrote:
> "Marshall" <marshall.spi..._at_gmail.com> wrote in message
>
> > Probably because there are many interesting things to learn
> > from considering view updates as being related to solving
> > systems of equations.
>
> What things, exactly?  I don't see any correlation between
> view updates and solving systems of equations.

I'm guessing you mean you don't see the value. If you really don't see how the two relate at all, then you're in no position to critique the idea.

If you want to see the value of the idea, the best way would be to compare the results given by the solving approach with any of various other approaches. The other approaches all seem to me to be ad hoc, whereas the solving approach gives a very clear theoretical foundation.

As to the particulars, the idea isn't developed yet, so the detained comparison will have to wait.

> > For example, an insert into a union
> > view involving two tables is not deterministic because there are three
> > different combinations of values that could result in the same value for
> > the view.

Um. It appears you're not using the word "deterministic" in the way I would expect.

Again, not every system of equations has a unique solution, and I would expect the system to reject those that didn't.

> > That would be one way to do it, but you are the only one talking about
> > doing it that way, and it is, as you say, a poor choice of how to do
> > it. This is what's called a strawman argument.
>
> Huh?  Are you saying that in a view A(X) union B(X), inserting a row (X=22)
> into to view is somehow deterministic?

It is deterministic if the semantics of the system call for it to be deterministic. In my case, I would design the system to deterministically
reject such an update. (You might want to review the definition of "deterministic"; it does not mean the same thing as "uniquely specified.")

It is only a strawman assuming you were using a standard definition of "deterministic"; it's not if you were using it to mean "uniquely specified."

> > Some systems of equations are overspecified, some are underspecified,
> > and some are uniquely specified. Which one of those do you think
> > might be a good candidate for the kind of view update that
> > would succeed? Would fail?
>
> I don't think systems of equations apply to view updates, because view
> updates involve more than one state of affairs, the state before the update
> and the state after.  Systems of equations are either independent of state
> or involve one and only one state, so I don't think they are even relevant
> to the problem of view updates.

To me, the question is, do my ad hoc view update rules give better results, or does my equation solving approach work better? My expectation is that the latter is true, but as I said, I haven't worked out all the detail yet.

Marshall Received on Mon Aug 31 2009 - 16:49:00 CEST

Original text of this message