 |
Home > Section 7 > Subsection 7 (Normalisation - part 1) |    |
Learning outcomes for this page:
Be aware that 'knowledge'
of the data is required to be able to carry out Normalisation
Normalisation, the very name puts fear into some database personnel, has been cloaked in technical complexity. Unfortunately this is due to a simple home truth that most fail to admit:
"It [normalisation] was applied by DP [data processing] staff, but often with great difficulty as it depends upon a detailed understanding of the data used by the organisation, and such an understanding is rarely held by DP staff. Rather, it is normally held by managers and users who work with the data on a day-to -day basis.
To them, normalisation represents 'common sense'. " (Finkelstein 1990 p94)
Exactly what is meant by ‘understanding’ in this context is rather difficult to define. As you work through this section you will realise that it does not mean knowledge of a particular data type required for a particular field (e.g. integer or real) but is more akin to understanding that you must possess a hospital registration number before you can be booked into a clinic appointment. Admittedly this is rather a oversimplification but it does give a flavour of what I’m trying to say.
The people who work with the data on a day to day basis can be considered to be domain experts, and once taught the basics of data modelling (such as object modelling introduced earlier) produce data models that are intuitively highly normalised. They apply the normalisation rules without knowing it! Unfortunately there is no guarantee that this will happen and it is necessary for them to learn the basics of normalisation.
Because data models often produced in this way are normalised the actual process of applying the normalisation rules post hoc is frequently felt to be superfluous if not simply a waste of time, however it would be more appropriate for the post hoc examination of the data model for ‘normality’ to be seen as a very important quality assurance measure.
Key point: To be able to carry out normalisation you must understand the data
Luckily Finkelstein, 1990 has eased the pain significantly by developing a user friendly version of the process of normalisation which he calls business normalisation. We will be using this variety for the first few stages. The first stage of normalisation is now described on the next page.
For: Clinicians | NHS managers | Non healthcare workers
Written by Robin Beaumont & Chris Noden e-mail:robin@ieg-net.co.uk or chrisnoden@hotmail.com