Re: Serializability of Transactions and Automatic (Number) Generators

From: Paul <paul_at_test.com>
Date: Sat, 04 Dec 2004 23:17:34 +0000
Message-ID: <41b2458e$0$29770$ed2e19e4_at_ptn-nntp-reader04.plus.net>


Jan Hidders wrote:

>> This suggests that there might be some value in a data type for abstract
>> identifiers. It is a little weird that we use integers for keys so often,
>> when there is nothing otherwise integer-like going on.

>
> Indeed. The integers are in many ways just a clumsy hack and it is at
> least conceptually more correct to model them as abstract identifiers.
> What is chosen to represent the identifiers is an implementation detail
> and has no place in the logical model. In fact, this was already
> recognized earlier on by the theorists and there has been some very
> interesting research on that, but at the time this was sold under the
> flag of "object oriented" and sometimes called "object identifiers" so
> the relational fundamentalist are still very uneasy about it.

Wouldn't any abstract identifier stored on a computer end up with a default ordering in the minds of users though? Because underneath, it will be always be stored as numbers (e.g. the ASCII codes).

How would you have get a totally abstract identifier that has no ordering baggage coming along with it?

Paul. Received on Sun Dec 05 2004 - 00:17:34 CET

Original text of this message