Re: Date is Incomplete - database application software and database theory

"--CELKO--" <jcelko212_at_earthlink.net> wrote in message news:18c7b3c2.0405131142.2739cfa8_at_posting.google.com...
> >> This is my point Celko. The model is incomplete, and was that way
> since the beginning, at least since 1980 when people started using
> relational database products. <<
>
> I am not sure what you mean by "incomplete"; Technically, I can show

Technically, I mean there is a difference between the relational model as a database, and the relational model as part of a functioning theoretical treatment of database systems, especially the 25 y/o RDBMS software, and the emergence of the e-organisation. (See below)

> that the SQL model is equivalent to primitive recursive functions and
> that means it has at least the same computational power as a simple
> declarative language on a single stack machine.

It has far more power because it resides in the same software environment as the data itself. All other solutions are hacks. ;-)

> This is of no use in
> the real world, since nobody would do an implementation based on it.

That's not true. I have pioneered a method of that uses a simple generic (extremely tiny footprint) front-end/portal as the user interface to 100% TSQL applications, the dynamic and organisational sensitive user menu itself, being a TSQL stored procedure.

There is an engineering prototype that a few people over here are using which is emminently adaptable to the real world.

A mock demo is the southwind database (of TSQL), and it compliments the std northwind schema of SQL server: www.mountainman.com.au/software/southwind

> If you mean that we need to keep working on data theory, I strongly
> agree. There is not a good formal definition of OLAP as far as I
> know. Nor of reports. Data mining and Data Quality are still wide
> open subject areas.

This tool, which technically would be classified as a generic portal software for RDBMS software instalations, permits reports to be written entirely in TSQL. Export to excel, or print.

Its far more efficient than any 3rd party repor generator (eg: crystal reports) because everything is unified within one software environment, rather than scattered to the 4 corners of the planet within the (*dualistic*) client environment.

Data mining is inbuilt because 1,2,3,4 .... stored procedures can be chained together with really easy coding.

You want to design and implement a four layered data-drillable structure by writing only 4 stored procedures? I did! That's why I built the tool.

> Temporal and spatial data need work.

We need to work on a theory of organisational intelligence rather than stopping with the relational model of database systems from the perspective of the database systems.

> >> Alot of people think that because I am criticising Date I do not
> appreciate the work he has done. This is baloney. <<
>
> I have it worse than you do; I actually Chris and have found him to be
> a gentleman and quite funny in person. His keynote speech at DAMA
> last week was great.

I am sure he tries to enjoy the circus. But it seems to me that the further work yet remains.

> >> I simply ask: "Have we reached the ultimate level of understanding
> concerning the theory and management of relational database
> technology, as elucidated and published by Date et al."? Of course we
> have not. <<
>
> "The road goes on forever and the story never ends!"

"The Road goes ever ever on
 Down from the door where it began ..."

[Tolkien]!

Pete Brown
Falls Creek
Oz

"A day will come at last when I
Shall take the hidden paths that run
West of the Moon, East of the Sun." Received on Fri May 14 2004 - 11:30:56 CEST