Re: temporary databases
Date: 27 May 2002 11:17:03 -0700
Message-ID: <bdf69bdf.0205271017.3644dd34_at_posting.google.com>
Ruud de Koter <ruud_dekoter_at_hp.com> wrote in message news:<3CF1E5B7.BE02F58C_at_hp.com>...
> Never thought about the legal implications of knowing something at a given
> moment? Suppose someone is going to sue *you* for not saying that Exa owned the
> flat on Jan 7. It would be very interesting to know what you exactly knew at the
> moment you made the statement.
OK, if lawyers are involved I give up. However, letting layers to define database model maybe is not such a good idea.
> Still think this is a theoretical situation? Go talk with a real customer in the
> insurance business, or stock exchange.
I assume that the legal implications could be triggered by wrong transactions only. After discovering wrong transactions, a business runs compensating transactions in order to prevent any legal actions. Compensating transaction might be as simple as just returning customer's credit. Compensating transaction can refer to wrong transaction, so that a business would still be able to query transaction history for errors, or what did they know at the moment, if you like.
Anyway, wrong transactions are supposed to be exception, rather than
norm. Therefore, "transaction time" is just unsound abstraction
level. We learned from transaction theory that it's a good idea to
eliminate combinatorial state
"Transaction time" is increasing modelling complexity level instead of
decreasing it, because you consider much more states.
explosion when transaction participants are allowed to see
each-other's intermediate states (no isolation).
I forwarded your suggestion to Samuel L. Clemens. Received on Mon May 27 2002 - 20:17:03 CEST