Memories, we know, are fallible, and in the case of acts such as this week's bombings in Boston, this presents problems. How can those charged with gathering eyewitness accounts - and those charged with giving them - be sure of what they're hearing and saying? In a couple of words, they can't.
Last night, I tried to remember what I'd had for breakfast. A recollection of porridge and toast swam to the surface of my mind after a second or two, but now, less than 30 hours later, my memory is fading further - what did I have on my toast? - and I doubt I would remember anything even this specific if I hadn't gone through the recall exercise last night.
Only a few hours separated my breakfast from the twin explosions in Boston. But similarly, witnesses to that terrifying event are already struggling to retrieve the details of their memories.
Critical evidence that might lead to the arrests and prosecution of the perpetrators is disappearing.
But my breakfast and the bombings aren't the same, surely. Those of us who were alive at the time remember where we were when we heard the Twin Towers had been attacked, when Princess Diana died in Paris, when JFK was shot ...
We live with the assumption that certain memories, by the gravity of the events contained within them, will remain fixed for ever in our minds. But sadly research suggests otherwise.
The really important information needed by police investigators and the FBI doesn't relate to the moment of the explosion but rather to the preceding minutes and hours - a glimpse of a suspicious action, a raised voice, an unusual smell, a hurried movement out of place in the crowd, a car registration, an item of clothing, and so on.
What may be a critically important description of a person or object related to this crime will need to be dredged up from several hours worth of mundane memories of standing, waiting, watching - or running, exhausted - at the end of yet another marathon.
The human capacity for accurate and reliable observation and recall is lamentably poor, and much entertainment can be derived from the various exercises in observation and eyewitness accuracy using videos designed for such classroom experiments.
Extensive research into false memories has demonstrated the fragility of even our capacity to remember what we remember. When it comes to encoding and retrieving memories, one thing is abundantly clear: memory fades quickly.
For police and other investigators, time is absolutely of the essence when questioning eyewitnesses to a suspicious incident.
It might seem eyewitnesses' memories would be strengthened, and their longevity preserved, by exposure to footage and the stories of other witnesses and journalists endlessly playing on a variety of formats. But again, the research suggests otherwise.
In fact, there is a grave danger of contamination of memories through interaction with these sources of (possibly questionable) information.
So what can be done? Ideally, the police - and a large team of skilled interviewers - would interview everyone immediately.
Each potential eyewitness would be sensitively interviewed in a quiet, well-appointed suite with access to refreshments, restrooms and childcare.
The investigators would all be fully briefed and would work in synchronisation. Every interview would be recorded and analysed, and the intelligence made available to every other investigator and dynamically cross-referenced to every other item of intelligence as it emerges.
Okay, but this is planet Earth, 2013, not a science-fiction film. The best we can hope for is that a team of investigators small enough to maintain investigative cohesion, but large enough to cover the ground will be deployed immediately following a briefing and will start interviewing potential witnesses.
We can hope they won't waste time with the wrong people; we can hope they'll be able to collate the data quickly; we can hope people will be contacted and interviewed before they start to forget critically important details.
We can hope that, under huge pressure to get results, investigators won't compromise the reliability of the information by using coercive or leading questions.
But we will probably settle for a lot less than this, because that's just the way things are, right?
If we can turn our attention to a project in England, it would seem there is another way.
The Greater Manchester Police have a trick up their organisational sleeve called the Self-Administered Interview (SAI) - a paper-based interviewing system designed for circumstances such as those of the Boston Marathon blasts.
Field trials started in the UK in 2009 to see if the SAI was beneficial to police investigating crimes with witnesses.
One case study showed completed SAIs from witnesses were comprehensive and detailed, containing useful information for legally proving the case against the defendants, and establishing the charge as a joint venture.
Furthermore, using the SAI enabled them to identify three further key witnesses who provided formal statements containing important information concerning actions leading up to the incident.
The SAI form provides instructions to eyewitnesses so they can record their account of the incident immediately.
The completed forms are returned to the police and a small team of investigators can process the resulting data, maintain consistent lines of investigation, identify the key witnesses and contact them for a formal interview.
The subsequent interview is informed by the information gathered from the SAI, and the result is a more reliable and detailed account than is at present possible using existing witness interviewing protocols.
Manchester's public appear to have embraced the system, judging by high rates of participation and the pages of detailed accounts that have been submitted in field trials.
Police are enthusiastic about the project and keen to keep working with academics to expand it further to a wider range of crime scenes.
Will the system be available to the Metropolitan Police in time for the upcoming London Marathon? It's likely it will, thanks to the high level of co-operation between police and researchers in the United Kingdom.
Clearly able to see the benefits of this project, the British Transport Police have already begun field trials in the SAI.
Police in Norway have started training in use of the SAI, and those in the Netherlands are soon to follow.
And in the case of Boston? In the time it has taken me to write this article, more memories have faded, witnesses have been exposed to a barrage of potentially contaminating information and evidence has been lost.
Georgina Heydon is a senior lecturer in criminal justice administration at RMIT University in Melbourne.