Re: A pk is *both* a physical and a logical object.

From: Brian Selzer <>
Date: Thu, 19 Jul 2007 13:39:02 GMT
Message-ID: <WPJni.27005$>

"David Cressey" <> wrote in message news:bVHni.12870$s25.12266_at_trndny04...
> "Brian Selzer" <> wrote in message
> news:jrHni.23333$
>> "Roy Hann" <specially_at_processed.almost.meat> wrote in message
>> > "Bob Badour" <> wrote in message
>> > news:469e067b$0$8844$
>> >> Roy Hann wrote:
>> >> Unless we are doing a simulation. In which case, we seek to
>> >> manipulate
>> >> in the safe knowledge that we end up with a representation of what we
>> >> think the real world would have been had it started with the same
>> >> boundary conditions.
>> >
>> > Fair enough, and the practical example would be when we are running
>> > through a test script.
>> >
>> >> The truth of the matter is worlds are meaningless to mathematical
>> >> abstractions.
>> >
>> > I'll take your word for it. But surely we are interested in a lot more
>> > than mere abstraction here? We are interested in finding *just* those
>> > abstractions whose behaviour is a good analogue to the real world. Or
> to
>> > put it differently, we are interested in how set theory and predicate
>> > logic can be *applied* to real world data management. It's the
>> > "applied-ness" that makes all the difference, and there is only one
> world
>> > where anything can be applied.
>> >
>> > Brian's assertion that some arbitrary update can cause a possible world
> to
>> > become an actual world is like something out of Gulliver's Travels.
>> >
>> The schema of a database describes a set of possible database values,
>> each
>> of which represents a complete description of a situation that may occur,
>> and the body of a database, the actual database value, represents a
> complete
>> description of the situation that is actually occuring. In this context,
> a
>> world is a situation, an instance of the universe--a particular set of
>> circumstances that can or do obtain. An update clearly designates which
> of
>> the possible database values is now the actual database value, and since
>> there is a bijective mapping between the set of all possible database
> values
>> and the set of all possible worlds, it stands to reason that an update
>> causes a possible world to become the actual world.
> In the databases I have worked on, the cause and effect is reversed from
> what you said above. It's a change of state of the actual world that
> causes
> an update to be made to the database.


> This process is not perfect, because it depends on the state of knowledge
> of
> those empowered to update the database, and also on them carrying out
> their
> assigned role regarding database updates.

I agree. I/O error is a fact of life. Received on Thu Jul 19 2007 - 15:39:02 CEST

Original text of this message