CTO Articles
Published in IT World
E-Business in the Enterprise – September 6, 2005
A database, by any other name
By Sean Mc Grath
Isn't it funny that some of the most common words we use have the largest number of interpretations? I remember reading somewhere that the word "set" has the largest, um, set of interpretations of any word in English.
In computing, there are few words more commonly employed than the word 'database'. For many years I thought I knew what that word meant, but now I'm not sure.
When faced with mounting doubts and uncertainties about a word, the coward in me likes to classify the doubts and uncertainties into different categories. It does not solve the problem but makes me feel as if I'm doing something worthwhile even if I'm not.
Rest easy that I do not intend to foist the full awfulness of my classification of the word 'database' on you. However, I would like to talk about two of the primary fractures in the overall fissure.
In the technically-oriented IT community, the dominant interpretation of the term 'database' is a system of storing information in fixed column tables. A subset of IT folk have a broader definition that incorporates exotica such as object-oriented databases or document-orientated databases. The boundaries between all these variations on a theme are very blurry. Endless fun can be had in hotel bars at conferences searching for clear blue water between these various interpretations.
Let us not go there today. Instead, let us segue to the tea garden across the street to where the 'suits' have gathered for a soiree. Amongst the majority of C-level executives what does the term 'database' mean? My experience has been that this community is not concerned with tables and objects and documents and the like. These words do not feature in their interpretation of the 'database' word at all.
In their place, you will find all the 'ilities' that make up so called non-functional requirements. Things like reliability, scalability, availability and maintainability. To this way of thinking, a database is an application that has all these 'ilities'. A storage system that does not have them (or perhaps does not have them in abundance), is merely a glorified file system.
I think this difference of emphasis is interesting and helps explain some Twilight Zone [1] moments I have had in meetings over the years where one part of the room is thinking tables/objects/documents and the other part of the room is thinking backups and scalability.
If you find yourself in a meeting about a new system only to find that the constituents appear to be divided by a common language, the word 'database' might be a good one to define early on in the conversation.