An effective (though uncommon) approach to database design

In database design a lot of effort is spent on ensuring that database designs are perfectly normalized and slated for great performance. But there is one aspect that, though even more important, tends to get a lot less attention: correctness. If a database is completely normalized and performs brilliantly, but won't store the data that the business needs, or allows modification that should be rejected as violating business rules - what use is it? In this precon, Hugo Kornelis presents a method that enables the database designer to find out exactly what the business needs are. Common pitfalls, such as miscommunications due to abstraction and jargon are avoided because the method relies completely on using concrete examples in the jargon of the business. The method presented tells you exactly what questions to ask, how to ask them, and how the answers to your question should be incorporated in your data model. At the end of this day, you'll be able to make a conceptual data model for every application, and to transform that data model into a (completely normalized) logical database design.