Warning: Creating default object from empty value in /home/johnnz/public_html/wp-content/themes/simplicity/functions/admin-hooks.php on line 160

Can You Use Normalisation in SOA?

In many articles lately people have tended to use the term “normalisation” when they are talking about removing duplication from SOA (Service-oriented architecture).  This is a mistake.

The term “normalisation” comes from the world of Data Structure Modeling, where it describes a set of techniques for removing all unnecessary occurrences of attributes from the data entities in an enterprise data model.  It is NOT about removing duplication!

To use a term such as normalisation, which has a very specific meaning in one discipline, out of context (and, in this instance, incorrectly) in a related discipline is a very bad practice.  It is, sadly, a common error made by by many analysts and consultants.  They hear terms used in one discipline, usually those that are a bit “techy” or elitist, and then introduce them to another discipline.

In this new area, if people were being honest, the imported term has no real meaning. But, when one or two people begin to use it, with apparent authority, then many others join in.  Now, remember, the term in this new context, has no real meaning.  So, in effect, all of these people are talking “intelligently” to each other in meaningless terms and all believing that they know exactly what they are saying and hearing! It’s a little bit scary.

How do you solve the problem?

Option 1: If a word already exists in another discipline, whose meaning you fully understand and whose use in your discipline area would be exactly the same, then use use that word.

Option 2: If no word exists then use Plain English! If what you are talking about is removing duplication, then use the words “removing duplication”.


No comments yet.

Leave a Reply