Re: object oriented vs object relational

"James" <jraustin1_at_hotmail.com> wrote in message news:a6e74506.0206101136.25b5db6d_at_posting.google.com...
> > > > generic code to copy any data structure of arbitrary complexity
> > > > For details see: www.xdb1.com/HowTo/CopyObj.asp
> >
> > You've done a bang-up job showing how much more complex
> > and difficult your product is compared to an RDBMS.
>
> How did you arrive to the conclusion that the ability to copy any
> abitrary data structure is more complex in XDb than in today's typical
> rdbs without having provided a similar solution to compare with?

I already provided that, James:

Update some_relation
Set some_attribute = some_arbitrarily_complex_data_structure ;

> Wouldn't the fact that no one (that I know of) has yet provided such a
> basic operation TEND to indicate that the complexity of implementing
> such is actually more complex in typical rdbs?

Apparently, you don't get out enough.

> Both XDb and rdbs are fundamentally based on the same concept of
> relation (a related set of information, according to one book).

Well, Microsoft will tell you that the relational model is named after referential integrity but that don't make it so.

XDb is based on domains -- not relations. Weren't you listening before?

> At the most fundamental level, XDb and rdbs implement the same
> relation in a similar manner.

Unfortunately, you are sadly mistaken. An rdbms represents the logical interface to the user in a much simpler manner delivering physical and logical independence. Your product delivers neither type of independence and presents needless complexity to the user without any compensating benefit.

> Why is orthogonality of importance?

Orthogonality is an important concept, but you have not demonstrated any comprehension of it. You have only demonstrated that your product exposes needless structural complexity to the user -- without any compensating benefit.

Of particular importance to data modelling for databases is the "Principle of Orthogonal Design" by Date and McGoveran. Of course, that is only one place where the concept pops up. The higher normal forms, for instance, suggest that the components of a compound candidate key should be orthogonal. Received on Tue Jun 11 2002 - 02:33:02 CEST