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

On Mon, 31 May 2004 12:47:10 +0200, mAsterdam <mAsterdam_at_vrijdag.org> wrote:

>> Simplicity
>
>This reduces the statement to
>"It was mathematically proven that it is simpler
>than the graph based approaches." and leaves the
>judgement to the reader/student. An improvement,
>but it still leaves the questions unanswered:
>simpler at what? etc.

At number of instructions. Simpler in orders of magnitude, and susceptible of many automatic optimizations.

The superiority is very striking and overwhelming. That's why teachers don't spend a lot of time with this topic.

>While this does give some insights in the
>history of the use of 'data model' and
>related terms (for the people here who
>showed interest in that topic), it doesn't
>at all claim to mathematically prove anything.

Because you are quoting the wrong parts.

What about this?:

<quote>

CODASYL relational
GO TO 15 0
PERFORM UNTIL 1 0
currency indicators 10 0
IF 12 0
FIND 9 0
GET 4 1
STORE / PUT 2 1
MODIFY 1 0
MOVE CURRENCY 4 0
other MOVEs 9 1
SUPPRESS CURRENCY 4 0
total statements > 60 3

The relative simplicity of the relational solution is very striking. Note: In fact, the relational solution could have been reduced to just a single statement, a PUT; the GET and MOVE aren't strictly necessary. What's more (although Codd doesn't mention the fact), the CODASYL "solution" -- which was taken from another source, by the way, not created by Codd himself -- included at least two bugs!

</quote>

>But what exactly is compared? Relational model versus network model for
>interactive support(1) for nonprogrammers. To dissmiss all graph based
>approaches for all purposes based on it is overstretching it, IMO,
>jumping to conclusions.

There is only one graph based approach. The hierarchical approach is only a specialization of the network approach.

>Equating network approaches to graph based approaches, for all
>purposes?

If you want to formalize the network and hierarchical approaches the only way is graph theory.

> The network approach is Codd's formalization of the CODASYL
>specification for the purpose of interactive support(1) for
>nonprogrammers, in the documents you referenced.
>(Or should I say pointed me to :-)

Yes.

>To determine wether it possible generalise Codd's comparisons
>to relational approach vs. graph based approach, some
>more levelling is needed.

No, CODASYL has all the essential features of the network approach.

>I suspect we will not be able to agree on this one.

Surely. Rationalism is the contrary of faith. Faith means to believe without reason.

>However, maybe we can try to agree on the
>'mathematical proof' issue, by clearly
>stating what exactly was proven.

I recoment you to read Relational Database: Selected Writings.

Regards
  Alfredo Received on Mon May 31 2004 - 16:55:04 CEST