* show that it is typically used with sorted data,
* highlight commonalities with SQL's groupby and Unix's uniq,
* demonstrate valid uses for the default identity function,
* add some excitement by suggesting the range of possibilities.