Re: Extending my question. Was: The relational model and relational

From: Bob Badour <>
Date: Mon, 24 Feb 2003 01:40:12 -0500
Message-ID: <2wj6a.267$>

"Bernard Peek" <> wrote in message
> In message <1FX5a.233$>, Bob Badour
> <> writes
> >"Bernard Peek" <> wrote in message
> >
> >> In message <7Ik5a.164$>, Bob Badour
> >> <> writes
> >>
> >> >> You and Date insist that cardinality can only be determined by
> >counting,
> >> >> which
> >> >> requires distinguishability. I have a scale that allows me to
> >determine
> >> >> the cardinality
> >> >> of a multiset of tuna cans by their collective weight.
> >> >
> >> >If you cannot identify the cans of tuna, how do you place them on the
> >scale?
> >>
> >> You pick one of them up (any one will do) and put it on the scale.
> >
> >Exactly! And by picking one you have identified it.


> Nope. You have one can but its presence in your hand is not an intrinsic
> property of the can and so does not identify it.

It does if it is in your hand. The phrase, "the can in my hand", identifies a can. Date only claimed that a thing must be identifiable to count it. Cans on shelves are identifiable to human beings, which is why human beings can count them.

> The can you hold and
> any of the others could be exchanged without affecting any of the data
> in the system.

It would change the position of two cans, which are part of the system under discusssion.

> >>You
> >> don't need to identify which can you are choosing (given the assumption
> >> that all of the cans have the same weight).
> >
> >Ah, but the very action of picking one can up identifies the can. It
> >the can in my hand and then the can on the scale. It becomes identified
> >separate and different from all other cans.


> Nope. Picking up the can does not change any of its intrinsic
> attributes.

Identification never changes intrinsic properties. Whether someone identifies me as "Bob" or by my driver's licence # or as "that so and so on comp.databases.theory" has no effect on any of my intrinsic properties. In the example above, the can becomes identified as different from all the cans not in my hand.

> >If the cans were not identifiable, you would not be able to knowingly
> >one or to knowingly grasp one.


> You would not be able to knowingly pick up a specific can. But there is
> no requirement to be able to do that.

If you cannot identify cans, how would you even know you picked a can up? Whether any requirement exists is irrelevant if you lack any ability to identify cans. The statement implies a lack of sensory ability.

For counting cans, there is a requirement and luckily the person taking inventory generally has the sensory ability to identify cans, which is all that's required.

> >> >And before you can calculate a count from a weight, you must first
> >> >identifiable cans then weigh them.
> >>
> >> No, you have to count cans. You do not have to count identifiable cans.
> >
> >How do you know whether you are weighing cans if you cannot identify any?


> You can identify something as a can, without needing to identify it as a
> specific can.

Again, you cannot count it unless you specifically identify it as a can you have yet to count.

> >For the time the cans are on the scale, they are identifiable and are
> >distinguishable from all other cans both on and off the scale.


> If you take some cans from the multiset and place some of them on the
> scale you have split one multiset into two multisets.

How do you do that if you cannot identify the components of the original multiset?

> If you drop all of
> the cans into a bag after you have finsished weighing some of them, how
> do you pick out the cans you claim to have identified?

Apparently, you want to have your cake and eat it to.

By dropping the cans into the bag, you have made the conscious decision that you no longer need to know which can is which. After all, once you know the count, you may no longer need to identify the cans but only the bag they are in. If you did need to tell which can was which long after the count, you could make a conscious decision not to destroy the information.

By eating the cake, I make the decision that I no longer need to have a cake.

> >If you cannot identify cans, how do you know which cans you have already
> >counted and which cans remain to count?
> At no point do you need to identify any specific cans.

Yes, you do. If you have counted two cans, you need to identify that the next can to count is not either of those two cans. That requires identification.

> You can place the
> cans you have weighed into one pile

Which identifies them by position.

> and the ones you haven't into
> another.

Which identifies them by position too.

> What this does is to create a new attribute that didn't exist
> before you sorted the cans.

I disagree. The position attribute existed prior to sorting the cans. The value of the position attribute changed, that's all.

> You could just as easily take a sharpie and
> write a Can Number on each one.

What relevance does the above statement have? Received on Mon Feb 24 2003 - 07:40:12 CET

Original text of this message