i barry rosenfeld, i'd like to start right there.t's the problem that we need to understand in trying to determine in advance who might be capable of violence as we saw in new toub? >> well, the essence of the problem is that it's a needle in a stay stack. so we've got almost an infinite number of people-- i shouldn't say infinite-- a very large number of people who will fit any profile we might generate and we want to find the one person who's potentially going to be homicidal. there just isn't really a way statistically to identify or clinically to identify that person with any real accuracy. >> brown: dr. joshi, does that mean such limits we can't know what can be done? >> the issue, however, is that young children and adolescents who sometimes will have aggressive behaviors early on and i think whole issue of trying to access care early on would go a long way in trying to prevent some of the more aggressive and violent behaviors as these youngsters get older. so i applaud president obama's recommendations and proposals he's put f