Re: MV Keys (was: Key attributes with list values)

From: David Cressey <>
Date: Mon, 27 Feb 2006 11:56:16 GMT
Message-ID: <ANBMf.1658$SJ2.1262_at_trndny01>

"Marshall Spight" <> wrote

> I think that sets and multisets can both be handled well with
> relations, as can orderings if necessary. But I long for
> the day when mathematicians will develop a unified model
> that will handle both lists and peruvian monkey fish.
> Tasty, tasty peruvian monkey fish. Mmmmmm.
Odd. I lived in Peru for two years. I've tasted everything from caviche to anticuchos and papa a la huancaina, but I never tasted this. (I apologize to the group if I just hit "send" with no message)

BTW, I no longer long for the day when there will be a mathematical "theory of everything". Maybe it's an age thing. I expect to enjoy my declining years better than I can by longing for something that will probably not arrive in time.

As far as "odering" is concerned, there is an ordered set: the natural numbers.

And every finite set can be mapped onto a finite set of natural numbers, by just "doin' what comes naturally" (apologies for the pun). So if you represent sets of tuples, representing lists is essentially a trivial addition to the representative capability of the system.

What interesting here is how you convey intent. If you order a set for the sake of computational convenience, and not for the sake of content, that fact needs to be conveyed to the users of the data. If, on the other hand, you order a set for the sake of content, where the order conveys some substantive information, that fact needs to be conveyed to the users of the data.

I was about to call the users of the data the "data archeologists of the future". But then I recalled my own experience in the field. I've felt like a data archeologist making sense out of data that was barely a year old. Received on Mon Feb 27 2006 - 12:56:16 CET

Original text of this message