Re: computational model of transactions

"JOG" <jog_at_cs.nott.ac.uk> wrote in message news:1154690573.601286.285520_at_i3g2000cwc.googlegroups.com...
> Erwin wrote:
>> > It is not always the case that if more than one actor is
>> > updating the same resource, that those updates must be serialized.
>>
>> I leave that one on your account.
>>
>> > To illustrate this, ignore the business rules in the above example.
>>
>> Ignoring business rules is not my idea of a good example. And
>> especially not this kind of example, since I've been a bank programmer
>> for 15 years. I can assure you no one in the bank business would
>> accept the kind of risky manipulations with account balances that you
>> propose.
>>
>> > The semantics of the update involve modification, not replacement
>>
>> You obviously see a difference between modification and replacement. I
>> don't. So please explain.
>
> I've noticed that this has cropped up a lot recently, where there
> exists a belief that one is changing part of a fact, as opposed to the
> truth that one is deleting a now incorrect fact, and inserting a brand
> new one. It appears a primary example where a common and understandable
> intuition is misleading.
>

No, the circumstances underpinning a fact have changed, and the database must be changed to reflect that. A true statement is either an axiom or reflects circumstances that exist in a real or conceptual frame of reference that is called the universe of discourse. Axioms are always true and can stand by themselves, otherwise they wouldn't be called axioms; therefore, the only statements in a database that can change are those that reflect circumstances that exist in the universe. There is no contest in the chicken/egg debate when it comes to which comes first: a database contains knowledge about the universe, so when a change occurs, it occurs first in the universe and then becomes known by the database. So if circumstances change in the universe such that a statement about something relevent to the discussion that was known to be true no longer is, then it follows that that statement must be altered to reflect the new circumstances. In the same way, if circumstances change such that something new becomes relevant, then new statements must be added to reflect that change. Similarly, if circumstances change such that something that was relevant no longer is, then the statements about that thing must be removed.

Relations are a sound mechanism to classify statements about things so that the discussion can be directed toward some goal. The predicate of a relational database not only classifies the statements contained within but also limits the scope of the discussion. The presence in a relation of a statement about something indicates that that thing is relevant to the discussion. For example, if a relation contains statements about the location of people, then if someone's location changes, then there must be corresponding statements in the database states preceding and succeeding the change. The absence of a statement in the succeeding database state implies that that person's location is no longer relevant; the presence of an additional statement implies that an additional person's location has become relevant. (This is why I argued in an earlier thread that statements in one database state must be able to be correllated with statements in the next: how else would you be able to tell what changed, what became interesting, and what should now be ignored.)

Thus, a statement that is no longer true must either be modified to reflect the new state of the object of the statement, or it must be removed to reflect that the information is no longer relevant.

There is a significant difference between modification and replacement.

>>
>> > the operation
>> > involved, addition, is communitive and associative.
>>
>> You mean "commuTAtive", of course, and furthermore that's completely
>> irrelevant. As for associativity : it is important to observe that
>> each transaction in this example does exactly one addition with exactly
>> two arguments. So unless you can think of a way for the system to
>> detect that multiple independent transactions are doing such a thing
>> (performing an associative operation), and then replace those multiple
>> independent operations with one single, transaction-surpassing,
>> operation that has the same result, associativity is also irrelevant.
>> If you cannot think of such a way for the system to detect this (I'm in
>> that case), you're stuck with doing multiple additions one-at-a-time,
>> and you're stuck with the fact that for the additions that are executed
>> second and third, one of those arguments should be the result of the
>> former. Therefore it is necessary that the transactions be serialized.
>> Otherwise it would mean a transaction is allowed to see uncommitted
>> results from another one.
>
Received on Fri Aug 04 2006 - 22:29:06 CEST