Re: Base Normal Form

Gene Wirchenko wrote:
<snip>
> >I don't say that ordering is significant AT ALL, but I do say that
> >using the term "relation" and then insisting that tuples be unordered
> ^^^^^^^^^^^^^^^^^^^
> No, not insisting that tuples be unordered, but that their order
> is irrelevant to the RM and that you can not count on any specific
> order.

Completely agree. Tuples are ordered (by def typically, although one could define the term otherwise) and their order is irrelevant to the RM.

> >is redefining the term, thereby muddying terminology and notation
> >unnecessarily, in my opinion.
>
> Different contexts, different definitions.
>
> I think there may a question of which tuples are being referred
> to.

any of 'em, but I was referring to "column ordering" and not row ordering. The RM operators need not act on the column ordering info, but it might be useful to define and order for printing out a data catalog or whatever.

[snip]

> The knowledge I prize the most is the knowledge of what I know
> and what I do not know.

and someone once said that the most dangerous knowledge is that of which we are certain, but wrong. Cheers! --dawn

>
> Sincerely,
>
> Gene Wirchenko
Received on Thu Jul 14 2005 - 02:10:51 CEST