Re: Temporal database - no end date

Marshall wrote:
> On Jan 22, 9:39 am, "DBMS_Plumber" <paul_geoffrey_br..._at_yahoo.com>
> wrote:
> > Volker Hetzer wrote:
> > > I'm not really sure what's wrong with it. All it shows is that you have
> > > chosen the wrong granularity for the job. Your "pick any" sentence simply
> > > states that for any real world scenario you can choose the wrong granularity
> > > if you insist.
> >
> > As opposed to a model where you don't have to pick a granularity at
> > all, and where there is not even the logical possibility of this
> > situation occuring.
>
> If you've explained how to do that, I've missed it. How can you
> not pick a granularity? The pigeonhole principle would seem to
> indicate that you must. I assume we're still talking about software
> on digital computers here.

 How can you not pick a granularity? Just don't. What does the 'pidgeonhole principle' have to do with anything?

 We are talking about a logical model.

 Over and over and over again on this newsgroup, we have labored to explain distinctions between logical data models, and physical implementations of logical models. What the DDL book describes is a logical model unmoored from ("useful", "practical") questions relating to actual programming. DD&L quite rightly eschew discussions of physical implementation, and instead develop a consistent model for storing and manipulating temporal phenomenon.

  Their model takes the notion of a 'time quanta' as a fundamental building block, and large chunks of their subsequent development relies on this notion of a 'time quanta'. (See their definition of PACK() and UNPACK() operators, for example.)

  Now, the alternative logical model doesn't have the notion of a 'time quanta' at all. Instead it relies on units, paired point values on a continuum, and an appropriate class of operations.

  The objection is that, while adopting the 'time quanta' simplifies the task that DD&L set for themselves, it creates other problems, one or two of which I have described. The conclusion of most folk who have looked at the question is to adopt the second approach. If you like, 'time quanta' are an unnecessary entity, and the simpler model, the one with fewer basic entities, is better. Received on Mon Jan 22 2007 - 20:16:32 CET