Re: A new proof of the superiority of set oriented approaches: numerical/time serie linear interpolation

"Jon Heggland" <jon.heggland_at_idi.ntnu.no> wrote in message news:f1c26e$9nk$2_at_orkan.itea.ntnu.no...
> Brian Selzer wrote:
>> I don't think so. D&D's interpretation of the model is great for dealing
>> with individual snapshots of data. If there were no need for transition
>> constraints, then their interpretation would be sufficient. But I
>> perceive
>> a need, and even D&D perceive that need: it's RM Very Strong Suggestion #
>> 4!
>
> And they show how to solve it, in the very same VSS. If I understand you
> correctly, you reject their scheme because you consider it insufficient
> in the face of key updates. Right?
>

Right.

> I don't think the semantics of key updates is as clear-cut as you think,
> though. It may seem intuitively obvious that when you update the sole
> key of a "Person" relvar, you are still talking about the same person
> (speaking loosely). But this is entity thinking, which is fuzzy, and
> dangerous to mix with relational theory. I am skeptical as to whether
> this notion of "sameness" can be formalised, and generalised to the less
> obvious cases.

I don't think the semantics of key updates is even relevant. On the other hand, I do think the semantics of update is relevant. For example, if Q is the set of facts corresponding to the targeted subset of the current relation, then an update is the sentence,

exists x in Q not x or T(x)

where T is a function that transforms each fact from Q.

So, when a user issues an update, the mapping between the elements in the current relation and those in the proposed relation is provided, because an update is existentially quantified.

It seems to me that if the user can target a particular tuple, and then specify different values for its attributes, then I should be able to define a constraint with the same granularity without regard to whether one of those attributes is part of a key. Whether or not a tuple represents some entity or not is irrelevant.

> --
> Jon
Received on Thu May 03 2007 - 12:36:30 CEST