Re: identifying entities across database updates (was: Is a function a relation?)

Brian Selzer wrote:

>> You and I, or even I and I, may very well disagree on
>> how to partition the universe into objects, depending on the purpose of
>> our description.
>
>But you cannot deny that there can be objects that can be located in time or
>space, because there is overwhelming empiracle evidence to support it. Just
>look out the window.

I don't deny that there are objects. But look at

  http://en.wikipedia.org/wiki/Window

How many windows are there in each of the pictures?

>>How many objects is a chair? How many objects is
>> the character A? Can it be green? Is the chemical element C an object?
>> Is course 430, Database Design and Administration at Moron University,
>> an object?
>
>Unlike you, I am not referring to an arbitrary partitioning.

It's only arbitrary in the Saussurean sense, i.e. established by convention among those who use the terms.

[...]

>> The reconciliation of different, overlapping database
>> schemas, even within the same organization, is an important practical
>> problem.
>
>So what? How does that bear on whether there can be things that can have a
>location in time or space?

What those things are is part of our conceptualization. Different choices of concepts => different things. But I agree that for most cases this isn't much of an issue.

>> In order to apply the relational model, we don't have to assume the
>> existence of objects at all. It's an extra step you choose to make.
>> Aren't you just complicating matters for ytourself, and for the
>> rest of us?
>
>You are so totally wrong: if we don't assume the existence of objects, then
>there is no need for the basic assumptions of the relational model: the
>unique name assumption, the closed world assumption, and the domain closure
>assumption.

These assumptions are useful and common, but optional.

[...]

>> I think that's a bit much to ask. You can't quantify over them
>> unless you know how to find them and to tell them apart.
>
>Actually, you've got it backwards: you can't find them or tell them apart
>unless you can quantify over them. Properties are expressed in logic as
>1-place relations that range over the objects in the Universe.

I think you've got that backwards. True, most mathematical descriptions I've seen reason about sets of objects. For instance, in the definition of state machines, we have a set of symbols and a set of states. But these are essentially placeholders. The definition doesn't describe or prescribe what symbols or states are, it only describes how they are combined into a state machine. Anything may take the role of a symbol or state, although the formulation suggests otherwise. There is no such thing as a universal division of objects into symbols and the non-symbols, or states and the non-states, independent of a particular formalization in which these terms are used, such as a particular formalization of state machines. What we can do is compare objects within an instance of a formalism, or across different, but comparable instances of the same formalisms or of comparable formalisms.

[...]

>> Existence is a property of predicates or properties, not of
>> objects or values.
>
>Neither properties nor predicates can have a location in time or space;
>therefore, you're wrong: existence is not a property of predicates or
>properties. It is nonsensical to state that red exists, but one can
>sensibly state that something that does exist is red. "The balloon is red."

You're right, I didn't formulate that correctly. We can't say that red exists, but we can say that redness exists: it means that some things are red. That is what I meant to say.

[...]

>> And this is your weak point. You want to express identity beyond
>> identifiability by attributes,
>
>This is not true. I didn't say this, nor did I imply it. I wouldn't even
>think of it because identity is a relation, and has a strict formal
>definition both for ordinary objects and for abstract objects.

OK. I don't think your formal definition completely succeeds but I see what it tries to do. What I disagree with is your insistence on objects being universal and absolute.

>You are
>conflating the concept of identity with the concept of identification.

Yes. I don't accepting the difference you claim there to be.

>They
>under the false notion that what identifies something at one time identifies
>the same thing at all times. While that is true of abstract objects since
>abstract objects cannot have a location in time, it is clearly not true for
>ordinary objects, as I have pointed out on numerous occasions.

For most concrete objects, most of the time we mostly agree on what they are. To that extent you're right. See the window example above.

>I guess this is partly my fault: I should have stuck with the word 'thing'
>rather than 'object' due to all of the OO baggage that 'object' carries with
>it, but for the record: you brought up OO. I didn't!

I don't think there's been a confusion this time.

><snipped fallacious argument based upon the misrepresentation of my
>position>

That argument was the whole point of my posting. The rest was just an introductory diversion. I really don't care whether your metaphysical universe contains absolute objects as long as you don't confuse them with what we can actually identify. And that is what you appear to be doing with your argument about the need to separate updates from delete/insert combinations.

-- 
Reinier