Re: MV Keys

From: Marshall Spight <marshall.spight_at_gmail.com>
Date: 26 Feb 2006 16:41:48 -0800
Message-ID: <1141000908.277461.101140_at_t39g2000cwt.googlegroups.com>


dawn wrote:
> The other difference is that you don't go more than a couple of levels
> deep in the nesting in MV compared to the possibilities for nesting in
> XML. Cache' (MUMPS) has more nesting. I have never encountered a need
> for it and it seems to help with conceptual simplicity not to have more
> nesting available within a single file.

This is exactly what I'd expect.

If pure 1NF were always the exactly right thing to do, we wouldn't have varchar. Instead we'd just use another relation, with tuples like (foreign key, position, character). Would anyone like that? No!

(The canonical objection here would be performance. For the thought experiment, just imagine it's free.)

Instead of doing this, we allow attributes of list type. List! Special cased for character lists only, though, perhaps so we can better convince ourselves that we didn't just throw in a non-1NF construct. :-)

So let's say we relax a little and allow general list and relation valued attributes. Then we could move away from 1NF as far as we wanted.

How far might we go? I expect that what we'll find is that the sweet spot is pretty darn close to 1NF.

Marshall Received on Mon Feb 27 2006 - 01:41:48 CET

Original text of this message