Hi guys...I'm new here.Nice to meet you all
So,i've just started a lesson in the university i'm studying which called database systems. I have an exersise to do for relational & e-r model. I need your opinions (cause as i can see it's not something fixed)for some questions.

1)Which is the strongest spots of each model?
2)Is there any weak spot in each one?If yes, which one (maybe more than one)?
3)Try to imagine a model which concludes(completely or partly) both models.Does this make sense?If yes,what could it be this super-model?

I would appreciate very much if you could answer me these questions.
Thanks in advance!

1)Which is the strongest spots of each model?

relational model and e-r model are apples and oranges

the relational model's main strength is that it is based on math

the e-r model's main strength is that it is visual


2)Is there any weak spot in each one?If yes, which one (maybe more than one)?

relational model has no weaknesses -- some people will argue this point, but they are wrong!!!

e-r model's main weakness is that it does not always mean you get a good database out of the design effort


3)Try to imagine a model which concludes(completely or partly) both models.Does this make sense?If yes,what could it be this super-model?

this doesn't make sense

First of all thanks for answering.

-Can you please explain what do you mean when you say apples and oranges?
-Why it doesn't make sense (the super model)?Can you tell me some more things about that?
-And what do you mean by "out of the design effort"?

Sorry for all these questions but i'm completely new!
I hope i'll get an answer.

Tnx again.

Would you settle for no *instrinsic* weaknesses?

First off, it's not true. You *can* compare apples and oranges.

As to what it means... Say you're a judge whose job is to determine who has the best produce. If you choose some apples from Farmer Abe and some oranges from Farmer Bob your results won't make sense because the taste is fundamentally different.

Since everything you can do with the E-R model can be accomplished relationally, the relational model would be your super model.

If I were designing a toolset to do relational modelling, I would want to allow the user to use different strategies to draft his schema. In this case, the E-R model's structures would be defined relationally and you could use simple macro replacement to build the logically correct relational expressions.

If the user were a scientific user doing Matlab-type work, I would also want the vectors, matrices &c. to be defined relationally.

Thus, a vector variable v might be backed by a relational variable rv that has constraints on it that it have a PK column Offset and a column of arbitrary data Values. Vector operations would then be translated into the appropriate relational operations.

What's important is that it would all be application independent: having designed your structures in your Matlab-type program, you could browse them with a spreadsheet, and all the various apps could share the DBMS with stepping on each other's toes.

the relational model is not comparable to ER modelsthe ER (entity-relationship) model, on the other hand, is a way of drawing a diagram to show entities and their relationships