It seems as though as software developers mature they develop consistency in their approach to just about every aspect of their work, regardless if there is a good reason for adopting a particular practice or not.
For instance, in data modeling I developed the habit of always naming my tables in the plural – Employees instead of Employee, and such. There’s no reason for this convention, other than perhaps I copied what I saw from the Northwind database.
But it’s important to question these practices from time to time, and after over seven years of doing things the same way I have decided to make a change. And for the second time now (see my post The Importance of a Logical Data Model), it was a colleague: Steve Dempsey who initiated the change. So why would one opt for singular names over plural ones?
Developers might chose singular names because they are shorter and require less typing, but this argument never held for me because of tools like intellisense and code generation (not to mention touch typing). But Steve is extremely adamant about singular names for a different reason: because of relationship readability.
For instance, in Sharepoint, workflows relate to events. Specifically, a workflow (singular) is initiated by one and only one event, and an event (singular) can initiate multiple workflows, as is expressed below:
The objective of modeling is thus to express the relationship of a single entity (a workflow, an event, or whatever) to zero or one or many of another entity. So why not just name your entities appropriately in the first place: by making them singular?
Of course now the problem is getting an old dog to remember his new trick. Or is it tricks?