Re: Does Codd's view of a relational database differ from that ofDate&Darwin?[M.Gittens]

"Jan Hidders" <jan.hidders_at_REMOVETHIS.pandora.be> wrote in message news:eHrxe.135259$JD6.7251058_at_phobos.telenet-ops.be...
> vc wrote:
>>
>> What's the (a) formalism for the FDM comparable to the RM (the FDM
>> structure/the FDM algebra/FDM constraint definitions). Daplex does not
>> quite cut it as an 'algebra'.
>
> Why not? Because it has operations with side effects?

An algebra is a set S together with operations on S. An operation takes zero or more elements of S and returns an element of S. In other words, an operation (and the algebra) is 'closed' over S. If it's not closed, then it's not an algebra. Over *what* is an FDM algebra closed ?

>> Now, IFO, as far as I remember, deals only with the structural part,
>> namely, directed graphs representing 'atomic objects', 'constructed
>> objects', etc. No algebra/data manipulatation and no constraint
>> definition was suggested at the time.
>
> That was done in later work such as "Object identity as a query language
> primitive" by Abiteboul and Kanellakis.

Does 'the object identity as a query language primitive' qualify as an algebra (see above) and a constraint definition tool ? If so, could you provede examples of both ?

>
>> Do not know about the rest of the bunch but wonder whether the triple
>> <structure, algebra, constraints> has been defined for at least one of
>> the abbreviations cited.
>
> LDM (the Logical Data Model) had a logic defined with it.

I'll take a look at LDM later.

>
>> I tried to read one of the articles you've referenced
>> (http://www.orm.net/pdf/ER96.pdf, "Conceptual query language"), but
>> quickly stumbled over the words "semantic" and "conceptual".
>
> :-) Yeah, ORM people like to use those words a lot. I wouldn't worry too
> much about it. You can understand the query language w/o understanding
> what those words exactly mean.

Can your response be interpeted as meaning that the article is essentially non-sensical (we do not care what the words the authors use mean) ? I am genuinely puzzled. Without digesting the basic notions, like 'conceptual object type', one cannot move any further. Alternatively, if all you are saying is that I can safely skip words like 'conceptual' and 'semantic' for this specific article, I'll try to re-read it.

>
>>>No. In ORM NOLOTs are abstract. It is more correct to say that the RM is
>>>basically ORM restricted to LOTs. A very grave and crippling restriction
>>>indeed.
>>
>> What does NOLOT being 'abstract' mean ?
>
> That there is no value representation associated with non-lexical objects.

OK. Firstly, why do we care about 'value representation' at all ? Secondly, please point me to a more formal definition of, say, LOT. The reason I've been asking all the questions is that I am trying to see what the state of the art is with alternative (other than the RM) data models is and how useful they are or at least might be in comparison to the RM.

Thanks.

>
> -- Jan Hidders
Received on Sat Jul 02 2005 - 16:13:27 CEST