Thank you for that question, and thank you for paying attention to my now decades-old work.
I think representation and recursion are the two variables that I consider most important. One of the things we have discovered is that when scientists get together and decide what is relevant to their review processes, often a set of blinders descends. There's disciplinary narrowing of the field of vision of the review process. One of the ways in which the National Research Council has tried to get away from that kind of narrowing is to ask somebody to coordinate the review process who is not specifically a member of the group producing the results.
In broad public review processes, Parliament or Congress can serve that purpose, to some extent. As you know, all our agencies are accountable for showing up before relevant congressional committees. That breadth of representation, which, for example, in the Canadian context would take language sensitivities and minority position sensitivities into account, would be one way of ensuring representation.
The other point is that the perfect should not be made the enemy of the good. No review process will be guaranteed to produce truth positions. Having recurrent review processes that come back—for example, when there's a large centre grant, having it reviewed with site visits at various points during its operations—and keeping that in mind will take care of some of the temporality issues mentioned by Dr. O'Neil in terms of science's ability to deliver. It doesn't happen all of a sudden, where yesterday you didn't know the truth and tomorrow you know.