Dot Connecting

Children's books often include exercises to "connect the dots". If you connect them properly, a clear picture emerges. Of course, if the dots are connected improperly all you get is a meaningless jumble.
In the investigation of the Ft. Hood shootings, the question that keeps being repeated is "was there a failure to connect the dots?" (For this post, I'm ignoring the issue of terrorism vs. mass murder. The problem of 'dot connecting' is the same regardless of motivation.) Could the slaughter have been prevented? Were there enough warning signs that authorities should have taken action?
We go through this process whenever a massive tragedy occurs. After the Virginia Tech shootings, there were revelations that the shooter was deeply disturbed. The investigation of 9/11 revealed that there were many pieces of evidence that the plotters were working up to something big. In the aftermath of the Oklahoma City bombing reams of evidence came out that Timothy McVeigh had spent months building up to a massive anti-government act. In all these cases there were many pieces that in hindsight were "dots"; the question is whether they should have been "connected" in such a way as to reveal the picture that would prevent the ensuing tragedy.
During the years that I worked as a mental health counselor, one of the most important parts of my work was suicide prevention. The population of psychiatric clients includes a very high percentage of people at high risk for suicide. Sadly, despite our best efforts and 'expertise', some clients did commit suicide. When a suicide did occur, the treatment team would meet to review the case; to try to "connect the dots". While the goal was to hopefully find ways to prevent future suicides, there was the natural human tendency to assign blame: somebody failed to "connect the dots".
As I'm watching the process unfold (again) with the Ft. Hood shootings, I'm remembering how difficult it is to actually connect dots. In hindsight, here are a whole bunch of disturbing signs; do they add up to a clear picture of impending violence? Is there enough evidence to take action to intervene? Is the goal to prevent future incidents, or to blame somebody for the incident that occurred?
Example: You have a neighbor, "Mr. X"; he's anti-social, often angry. He blames "Y" for all that is wrong in his world. He might have weapons. What signs would be clear enough for you to call the police about "Mr. X"? How do you differentiate between an eccentric crank and a potential killer? Remember that our society respects an individuals freedom (in theory, anyway), and that "Mr. X" hasn't done anything yet. How many dots connect and what picture (if any) do they reveal?
Intervention is a very inexact "science"; you never really know what might have been "prevented" because there is no way to measure a "non-incident". Maybe a tragedy has been prevented, but maybe prejudice and hysteria have needlessly disrupted a persons life or deprived them of there individual freedom.
I don't have any answers; I hope the investigation will focus on trying to find ways to prevent future tragedies by helping people who are deeply disturbed, rather than degenerating into finding someone to blame. The reality is that life is full of uncertainties and that tragedies occur daily. Preventing atrocities is an admirable goal, but not all dots connect into a clear picture.