Re: A Normalization Question

> Brown is a string" is a true proposition.
> "Brown" is just a string, IOW,

I contend 'brown' is a fact. I think you will agree that 'brown is sequentially composed of b, r, o, w, n ' is a fact or true proposition. The reason 'brown' is a fact is because it is equivalent to 'brown is sequentially composed of b, r, o, w, n '.

RM can't determine 'brown' is a fact until, it is represented according to a special arrangement in attribute values of tuples.

You can't determine 'brown' is a fact until, it is represented according to a special arrangement called a sentence.

Both you and RM have limited definitions/rules.

> This gets back to what everyone else understands:
> Normalization takes place in a meaningful (E.g., business) context.

The above is vague because of the word "meaningful". Meaningful to whom? Something may be meaningful for one while not for another. Somethings are meaninful one day and not the next. The most general level of normalization takes place at the level of things, not tuples, attributes, values, lists, bags, tree, relations, etc.

> It is not about bits, bytes, strings, or physical storage.

While physical things can probably be normalized at the hardware layer, I am not talking about normalizing anything at the hardware layer when I normalize symbols and strings represented at at the logical layer. RM is a logical model. Implementation of it attempt to keep users in the logical layer. Data entered in Sql Server or Oracle is at the logical layer, and that includes symbols and strings. Why do you keep insisting that RM's implementations are allowing user to enter non-logical data? Please explain how to determine which layer user-entered data in a RM db belongs to?

Bits, bytes and strings can be represented at a logical layer, however they are not required at the hardware layer. For example, the human brain represents them without having them at the hardware layer. If one looks in the logical layer of an XDb1 db, one will not find bits, bytes or integers unless a user chooses to represent them. Received on Thu Jul 15 2004 - 20:48:57 CEST