At times we feel that we are getting everything right in one area of our work.  Everything we touch seems to turn to gold and we feel completely happy with our decisions and achievements – in that area.

But data modelling won’t let us work in only one area – ignoring all the bits we don’t like.  We can’t concentrate on just getting our data types right or making sure that we only model the details we need.

We need to put it all together at once: providing predictability, consistency, integrity, performance, completeness, simplicity and all of the other things we aim for in data modelling.

So how do we manage this?

Checklists

Checklists can help.  There are many around and they are often specific to a particular course or company.  You may need to start with a checklist which seems the most helpful and make small adjustments until it suits your requirements.

Larger organisations will often have much more detailed standards and checklists – many of them, in fact.  Smaller organisations may not bother with checklists at all.  I believe that checklists can be helpful in maintaining consistency and thoroughness.  They can help us avoid silly mistakes.

Some data modelling checklists are available on the web.  Here are a few simple examples from a quick search:

There are also some checklists included in books which are available for purchase.

Aside on Terminology

Normally, I try to use simple and general terminology.  I believe that if you can understand and describe data modelling in simple terminology, you will not find it hard to understand and learn the more specialised terms which often focus on particular tools or methods.

As a result, I tend to avoid the use of names like “Entity-Relationship Diagrams” (ERDs), “logical models”, “physical models” and similar things because they will often obscure the basics of data modelling.  Using such buzzwords can lead us into learning how to apply a particular method rather than understanding why we should choose such a method.  Once we understand modelling in general, gaining an understanding of the specific processes and terminology which are currently popular is easy, and can be very helpful – but an overall understanding of the field should be our first target.

I also avoid the use of diagrams with special symbology.  Since there are several different “standard” diagrams available, they can be learned once the basic ideas of data modelling are mastered.  Often an employer or customer will want you to use a particular tool – but these are easy to learn if you already understand the fundamentals.