Re: Database design

Gene Wirchenko <genew_at_ucantrade.com.NOTHERE> wrote:

>On Tue, 21 Feb 2006 18:14:12 -0800, Mark Johnson
><102334.12_at_compuserve.com> wrote:
>>"JOG" <jog_at_cs.nott.ac.uk> wrote:

>>>This whole 'flat' debate is nonsense too. Write a database down in its
>>>mathematical form, devoid of tables

>>It is flat. A relation has no structure, unless there is some ordering
>>among the attributes. The attributes are supposed to be in no
>>particular order. The entries/instances/tuples are supposed to be in
>>no particular order. That's just definitional.

> It is n-dimensional, one dimension per attribute. The three
>dimensions of Euclidean space are not in any particular order either.

But the order of finish, is. A relation of those in the race might suggest one, and not the other, won, placed and showed. But someone might object, that's not an ordering, but simply an unordered descriptive attribute. I would reply that it represents a proper order.

But I would further suggest another example, in that case, the one with which I began. A book. One paragraph does not go just anywhere. A title cannot appear on page 6 if it properly belongs on page 34. There's an ordering. There's a sort. Term it, partially ordered, if one prefers.

And at what point would a relation of paragraphs, say, which would include a sort attribute, have to be fairly termed an ordered relation? Or is such simply defined out of the realm of possibility, which makes it look like mere semantics, by the popular sense of the term? Received on Wed Feb 22 2006 - 08:16:50 CET