Re: The Theoretical Foundations of the Relational Model

On 26 Jun 2002 09:57:51 +0200, hidders_at_hcoss.uia.ac.be (Jan.Hidders) wrote:
>> Actually, no, I am constitutionally unable to "know" what the three
>> layer model is supposed to do for anybody, and this is a perfect
>> example why.
>
>Hm? There is a law in the US that forbids you to understand the ANSI/SPARC
>model? Or is it a health problem? :-)

It's a personal preference for not claiming to understand gibberish.

>> My question to you is why actual database engines and standards should
>> not allow ordering as part of a view.
>
>If you would have known the differences between internal schema, conceptual
>schema and external schema you would have known that I already gave an
>answer to that question. Views are in the external schema so ordering is
>allowed.

So, if a product is going to support the three-layer model, they would have two kinds of views, internal and external, and one would allow sort and the other would not?

Say, how about data warehouse products, that's where you'll find situations where physical data ordering makes more sense. And a rationale, even: databases "allow" you to physically (and logically) denormalize data, one can consider including sort on tables/views to be a denormalization, and actually a less harmful one than duplicating data.

--

If you have a reference you think makes the three-level model seem
sensible, please suggest it.  I'll take another look next time I'm at
the bookstore.  A quick Google around doesn't turn up anything I an
point at.  But my experience is that "conceptual" database models go
nowhere without extensive extra tools for use cases, dataflow, and
whatnot, "logical" models are insufficiently detailed for any kind of
planning or decision making, and only the "physical" model, based on
the actual DDL to create the database, is worth posting on the wall.

The three-level model *seems* like a good idea, and maybe it can be
salvaged, but the only encounters with it, or its advocates, I've ever
had have been massively unconvincing.

J.