Andres Aguiar's Weblog

Just My Code

Solving the Object -Relational impedance mismatch the other way

Ted Neward suggests solving the object-relational impedance mismatch by adding relational features to languages, instead of mapping relational databases to objects.

I like that.

 

 

Posted: Nov 13 2003, 10:45 AM by aaguiar | with 6 comment(s)

Comments

Thomas Tomiczek said:

Idiotic if you ask me.

I do objects because the OO paradigm is better than the relational one for developing software.

Adding relational stuff toa language would make reporing easier, granted, but it would not help at all with architecture.

Non-Issue, totally.
# November 13, 2003 9:02 AM

Andres Aguiar said:

>Idiotic if you ask me.

Did I ask you? ;)
# November 13, 2003 9:17 AM

Thomas Tomiczek said:

::Did I ask you? ;)

Yes, you did. You posted something and it asks for comments.

Naturally logical thinking is required to see this as an invitation.
# November 13, 2003 9:24 AM

Fabrice said:

I depends on who's invited, Thomas ;-)
# November 13, 2003 10:03 AM

Frans Bouma said:

It's not wise to do this as it will create the same type of relations as there are in the relational model: relations between types (vs relations between entity types)

A relation based on DATA INSIDE an object is what's needed and this is a SEMANTICAL relation, which means that the relation only exists for the object/functionality which interprets the data as a relation.

You can do that already, easily, so why should it be part of a language?
# November 13, 2003 10:27 AM

Andres Aguiar said:

Because objects impose constraints in the way I can use my data that don't exist in the relational model.

The same constraints that made hierarchical databases fail and relational database suceeed.

Your semantical relations are a subset of all the existing relations. If can have the full relation set available, why would I want a subset of them?

# November 13, 2003 10:39 AM