A view is just a virtual table presenting the result set of a a query--that's usually the way we perceive it in our everyday work, neglecting part of the power of views. Thus we use views to encapsulate complex queries, to subset or aggregate underlying tables, or to provide some additional access control to our data. However, a view can be seen as a fundamental building block of the declarative programming language that SQL is, such as a clause in Prolog or a function/procedure in an imperative programming language.

We'll discover a fun example that turns MySQL's command line interpreter into an expert system and allows us to have a chat with MySQL. It's done without a single line of imperative code (no stored procedures or functions), just by a clever combination of views.

This should give you an additional approach to solve some data mining problems "the SQL way" and to let the database itself make decisions based on your data. We'll see how this approach is put to practical use in a system for severe weather alerts based on a wealth of meteorological realtime data.

While you might not implement your next project using merely views, this session will definitely change the intermediate to experienced programmer's view on views.