It was about the efficiency.
That's where we were talking about using other information and other approaches so that the screening doesn't happen strictly at the checkpoint.
If it happens throughout the process and we have information fairly early on in the process, we can build systems. They're not existent now in the Canadian context, but it's possible to develop systems that will allow us to screen from earlier in the process than just at the checkpoint. Then we can stream people through—to the point that has been brought up by numerous members.
We're using the technology for those people who need it, not for everybody. So those delays are not felt by everybody, and therefore, the queues actually become shorter for everybody. The idea of having a random component is valuable, but relying on random as the only determinant factor as to whether you get that additional screening we feel could be improved upon and is not an optimal situation. So if you were to use more of the intelligence and the information that we have about people and what we can observe about their behaviour, you could stream them--high-risk, low-risk--and use those technologies more effectively.