My notes about interesting article “Enterprise Data Modeling: 7 Mistakes You Can’t Afford to Make”
About “Mistake 1: Forgetting that an enterprise architecture is a living framework”
An enterprise architecture is a living framework but it is not 100%-true.
“Living framework” in the projection to complement and minimal changes in the main part of schema (model).
If you have an application using the schema (er-schema) will be very expensive (time, cost) to change the application.
Talk about the life of the model (logical) is not correct in my view, correct to speak of the complement – the development of the model.
In my opinion development of the schema (in particular logical) is very important and requires the most expensive in the design phase, as further changes are overhead.
I think we should not talk about “an enterprise architecture”, we need to talk about data models and management of these models within the enterprise.
About “Mistake 2: Keeping data models invisible”
Not entirely clear what is meant by the term “definitions”.
“A model without definitions is just a diagram that could be interpreted in many ways…”. It’s not true.
If you have application which use er-model (for example HR) – how it will be interpreted in many ways? Not entirely clear =/
What is “invisible”? Invisible from whom? from the members of project? It isn’t true. Then he will not be able to develop applications =/
About “Mistake 3: Assuming that business users can’t understand or review models”
Who is “business users”?
Suppose we have a person(1) who develops an application and the person(2) who develops the database schema. Who’s Who?
If it’s =) person(1) then how he will develop an application if he does not understand what are some objects in the database (all the objects are not important to him).
If it’s =) person(2) then how he will develop the database schema? :-0
About “Mistake 4: Thinking that data models are only about databases”
Data models are not only a databases. It’s true.
Let’s see definitions of “Data model” on wiki.
About “Mistake 5: Throwing models “over the wall””
“A data modeler is the mediator between business requirements and physical implementations.”
Nice remark.
About “Mistake 6: Forgetting about the sizzle”
Not entirely clear.
In my opinion it is not as important (color, guides for nonmodeler) as the completeness and rationality of model.
About “Mistake 7: Thinking of them as “your” models”
Correct note.
“That means sharing them openly, providing access to those who want it,”.
Totally agree.
Thanks Karen Lopez, Kamille Nixon
Andyblg's Blog 