Re: Can we solve this -- NFNF and non-1NF at Loggerheads

"paul c" <toledobythesea_at_oohay.moc> wrote in message news:v07Od.329533$Xk.262055_at_pd7tw3no...
> Dan wrote:
>
> ...
>> Here we go:
>>
>> First Normal Form: A relvar is in 1NF if and only if, in every legal
>> value of that relvar, every tuple contains exactly one value for each
>> attribute (Date, 7th ed. p. 357).
> ...
<snip>
> apart from that, i always have a problem with the above "tension". can't
> quite swallow the word 'model'. after all, everybody knows that a
> computer doesn't duplicate reality, just represents some aspect of it.
> even if the computer reflects our meaning, how do we know that our
> intelligence models reality?

This is either a quote or variation on one from George Box "All models are flawed, but some are useful"

A model is like a metaphor -- it does not have to map every aspect of reality, just those that are useful for whatever you are trying to do. Relational theory gives one approach to modeling data and it has pros & cons compared to other models. I think someone was right on when they suggested that if Date & Darwin want to redefine 1NF, then they should call the new model by another name. Then we can compare the relational model to the-newfangled-relational-model to a variety of other models, looking at pros and cons related to the features desired for the project(s). However, relational theory has had a mind-share near-monopoly for so long, we need to get beyond that.

--dawn Received on Wed Feb 09 2005 - 01:34:06 CET