Knowing the interest measures of particular algorithms will allow us to protect against them, either by ensuring that the data we publish does not have ``interesting'' rules, or by providing data that gives ``false'' rules that hide the real ones. A standard specification of these rules would allow us to develop standard techniques for preventing mining.
Publishing details of algorithms, particularly how interest levels are determined and low-interest items are pruned, will enable countermeasures to those algorithms. We hope we will be able to find commonalities among algorithms, enabling us to develop strategies to defeat a wide variety of data mining tools.