Re: Data Modeling "the third way ?"

Dario Rossa wrote:
> Hi,
>
> Few days ago I head a discussion on data modeling.
> To me , at least what I know about data modeling there are two ways
> (or two families) to model "the real world" the data of your business
> model.
> 1. Dimensional modeling , Star Schema....
First order is Hierarchical Model (eg the unix file system) Second order is Network Model

> 2. Entity Relationship modeling , ERD , 3NF ..or E-ER modeling..say
> the OO way.
>
> But he had following modeling technique...
>
> - No relationships between tables.
> - Table A contains in one of the columns the name of the related table
> - Table A contains the record number of the related table, mentioned
> in the column above.

So actually the first step was a lie.

>
> It seems to me that this is not maintainable and to get or insert data
> you almost need to implement a database engine yourself to update
> record numbers...etc.
> No mapping possible to EJB CMP or whatever technology to retrieve or
> store data.
>
> His argument is : It is flexible...you can store everything...
>
> But what are you modeling....This is not a model to me and not usable.
>
> Happy to hear some comments on this one...."the third way"
>
> //The truth is out there
>
>
> Regards,
>
> Dario
> dario.rossa_at_mature.be

He is implementing a network model.

  The trouble is, if he implements this in ORACLE or other RDBMS he is going to have all the overhead of the network model and little of the benefit. A network model DB can be extremely fast on data retrieval. His approach won't be fast since he'll be building dynamic queries all the time. He has to maintain and follow the links himself for inserts and selects. Done in ORACLE or INFORMIX, the only benefit he bought this way is updates may be simpler then a pure network DB.

You are right that there are maintenance headaches with such an approach. I started out on a network model DB (HP's IMAGE 100 DB). Almost all of then get corrupt indices on occassion. It can be very messy to repair.

In case he thinks this is something new, let him down easy. The concept is _so_ 1960's.

At least he's not trying to sell it as "advanced technology". 8^)

-- 
Ed Prochak
running    http://www.faqs.org/faqs/running-faq/
netiquette http://www.psg.com/emily.html
--
"Two roads diverged in a wood and I
I took the one less travelled by
and that has made all the difference."
robert frost