Re: In an RDBMS, what does "Data" mean?

From: Eric Kaun <ekaun_at_yahoo.com>
Date: Wed, 09 Jun 2004 15:39:42 GMT
Message-ID: <2bGxc.6816$n03.2311_at_newssvr32.news.prodigy.com>


"Tony" <andrewst_at_onetel.net.uk> wrote in message news:c0e3f26e.0406080151.2f2f3221_at_posting.google.com...
> Where did you get that axiom from that "data comes in tuples"? Codd's
> rule #1 says that all data in the database is to be REPRESENTED in
> only one way: as values in attributes of tuples. It is a prescribed
> RULE for building relational databases, it is not a claim that
> anything in the real world "comes in tuples". We have a similar rule
> in English that all objects are represented by words made up from the
> 26 letters of the alphabet; it is not an "axiom" that says that
> objects "come in" combinations of the letters A-Z.

Ah, an excellent analogy. I'm sure it's flawed, but it gets the point across in a new way... thanks.

Another, often-cited, is the difference between "flat" tables and relations, and the way people assume relations are 2-dimensional. Consider the following:
0 0 0
0 0 1
0 1 0
0 1 1
1 0 0
1 0 1
1 1 0
1 1 1

Wow! A flat cube! Nifty! I was just too lazy to type out a tessaract (hypercube)...

> Your problem is that you consistently confuse data and reality.
>
> Of course, this all doesn't mean that tuples are the BEST way to
> represent data, or even that ALL data can be represented by tuples.

Keep in mind that even if you say this, the orthogonal dimension is Type (Domain), which introduces wrinkles of its own.

> But you could easily disprove a theorem that said that "all data can
> be represented by tuples" by finding a counter-example. Bet you can't
> though!

Yes, good point - find us that black swan.

  • erk
Received on Wed Jun 09 2004 - 17:39:42 CEST

Original text of this message