Re: Impossible Database Design?
Date: Sat, 20 May 2006 12:36:22 +0200
Message-ID: <446ef063$0$31648$e4fe514c_at_news.xs4all.nl>
-CELKO- wrote:
>>>Well, yes. As far as as I understand it, they (DD&L) discuss a discrete (vs continuus) universe representation - at least as far as time is concerned. Why the 'but'? <<
>
> I cannot shake Zeno's paradoxes which occur with a discrete model of
> time.
The computerized model by the academics you mentioned can? Which model? Who are they?
>>>I'm not sure if I am getting this the way you mean it. You mean p1, p2 etc. as time points? <<
>
>
> NO, P# as part numbers!
Part intervals do sound weird, I can't think of a sensible interpration right now.
I had to look it up. I didn't remember they used part-points on page 355.. (is that what you are talking about? - if not could you please provide I reference so I don't have to guess again?) in their example with two interval attributes in a relation.
Part intervals do sound weird. I read over it and at the time did not think it was important, because they just use it to demonstrate that the PACK and UNPACK operators not only work on the typical case where there is just one interval attribute, but still work on relations with more of those.
> He uses the [<start> : <terminal>] notation
> for anything
Well, anything as long as the requirements for
the Allen-operators on intervals are met.
One needs very specific parts for
that - maybe e.g. rail-sections?
>>>During Birds [p1:p5] pink lawn flamingoes <<
>
>
> Exactly! See what I mean about how it does not make sense! Then they
> have PACK() and UNPACK(), etc.
They use PACK and UNPACK to show that working with interval attributes does not compromize the relational model as they see it.
>>>I'm still wondering wether my interpretation of your questions is correct. <<
>
>
> Yes, and it really is that crazy for a set of unordered distinct
> elements whose names can be put into a sorted order.
I don't think that is what is going on in the examples. The 'names' you are referring to are, so it seems to me, just nicely chosen symbols in order to not confuse the reader.
> It crossses
> logical and physical bounds and makes no sense.
No, it is discrete - and you don't seem to accept that approach.
>>>Not as such, removing Tuesday from an interval over cyclic weekday points renders two weekday intervals. <<
>
>
> But I cannot remove Tuesday (the time period whose name is "Tuesday")
> from the cycle. About 40 years ago, the US had an animated cartoon on
> the Captain Kangaroo Show called "Tom Terrific"; in one episode, the
> villian found a way to remove the weekend so the kids would have to be
> school all the time and coudl not play. It was so absurd that young
> children could laugh at it.
>
> [Monday:Monday] ...
> [Wednesday:Thursday] ...
>
> Does not destroy [Tuesday: Tuesday] -- you mean that your entity was
> doing something we do not know about in that time slot. That model
> does not apply to Parts.
I'll give it a try with the rail segment interpretation. Here's a working schedule for our rail maintenance crew:
During Segments
[d1:d5] [p1:p4]
Interpretation: on day one to day five we'll do maintenance work on rail segments one to four.
Event: we got orders to repair segments six to nine on day two with all our capacity at the cost of postponing our work on [p1:p4] one day:
During Segments
[d1:d1] [p1:p4] [d2:d2] [p6:p9] [d3:d6] [p1:p4]
Event: Segment p3 appears in good shape, the work is finished after d1.
During Segments
[d1:d1] [p1:p4] [d2:d2] [p6:p9] [d3:d6] [p1:p2] [d3:d6] [p4:p4]
>>>Heh - it would become a strange week indeed. <<
It's a pity that you have a tendency to snip the comments I think are most relevant.
> The WB television network in the US did a series of advertisements for
> extra days of the week in which children could watch their cartoons.
>
> I granmt that we could l go back to the Roman and African ten-day
> cyclces, but this is just re-naming. In the DD&L model, I can actually
> create new temporal periods.
Received on Sat May 20 2006 - 12:36:22 CEST