As usual, what happens is an interim process. As we get more reports from the field that seem to match the same description, those are added to the reports that are sent to Japan. Each time one of those is triggered, that data is then put back through the system to engineering.
Engineering itself begins to look at the process. Typically where a part is developed with an outside supplier, they'll work with quality control at that supplier to try to understand what the problem is and replicate the condition. That replication was subsequently done. I can't give you a precise date for that, but I'm sure we can supply that to you later if that's necessary.
They look to then identify not only that they have identified one situation where it occurs but they've identified the only case where it occurs. They've put it through multiple testing in order to try to determine whether or not they've isolated the cause. In this case, they were able to identify a condition of both wear and condensation. As that happens, you then have to go about trying to engineer a solution to it. You can't trigger a response without being able to know that you have an engineering solution to the problem you've identified as well.
In that process, pedals are cut apart. They are subjected to various types of engineering treatment to try to determine whether you've fixed the problem and they are tested under more extreme conditions than you would normally encounter in the marketplace--hundreds of thousands of cycles on a machine in exposure to high condensation, high moisture conditions to ensure that not only can you not reproduce the condition immediately after you've made the pedal adjustment but that it goes through what amounts to many years of hard, in-service use. We have to replicate that in a lab. Obviously we can't have 15 years worth of experience on it because there wasn't 15 years of experience between the first report and the time we were able to take a recall.
Once we've identified the problem, once we think we have a solution in development, we still can't take that solution to Transport Canada until we're convinced that it's something we can deploy into the marketplace. So we trigger a recall, and in some cases in the industry we trigger a recall many months before there's a solution to the problem. In this case, because of the technical issues related to trying to pinpoint the problem, the engineering was basically running on a parallel track. So there was about a week time lag between the time we were able to speak to Transport Canada and say we'll issue the recall, and the time we came back to them with support from engineering to say we believe this is a viable response to the problem and it will fix the problem.