The actual change is that we're focusing more on high-risk areas. One of the advantages we have is that we've collected data over several years, and what that tells us is that in our very low-risk areas we have a very high percentage of conformity, meaning that there is a rate of very few errors.
What we see is that in higher-risk areas we want to pay more attention to them to ensure that we're testing pilots in such a fashion that we can catch those kinds of errors, look at any kinds of trends that are happening, and adjust in terms of the testing.