Good, solid article on the date dimension.

I love role-playing dimensions - they can expand business users' understanding of their data to points that they had never considered.

For example, one of the solutions I had implemented was a warehouse for a medical insurer. On the date dimension, I added an attribute called "Season", which would have members like "Summer", "Autumn", "Winter", and "Spring".

Initially they only wanted to track the seasonality of certain treatments e.g. do people visit GP's more in winter than in summer?

BUT

One day they were slicing and dicing, and suddenly they saw they could use the same date dimension to track policy movements - they could see e.g. that more clients would join in autumn (just before the winter) compared to other seasons... Suddenly they had another perspective on their data, and his new view would have an impact on their marketing campaigns...

The date dimension is an absolute treasure in BI solutions, and allows analysts to bring together data, and gain real insight, into the movements of their business