edg | Feb. 5th, 2005

When one is examining injuries - from an analysis point of view, not a medical point of view - there are two causes of injury that must be looked at: the immediate cause of injury, and the underlying cause of injury. The immediate cause of injury is the split-second event that directly caused the injury; the underlying cause of injury is the event that led to the immediate cause.

For example, let's say that a man is brought into the ER with a broken arm, having sustained this after falling off a ladder. In this case, the impact with the ground would be the immediate cause of injury, and the fall would be the underlying cause of injury.

I'm reminded of this because of a discussion on the Pyramid message boards, regarding the recent switch in log-in authentication schemes; it seems that certain web browsers no longer automatically remember the password. Some users recommended switching browsers; the initial complainant took issue with this, defending his use of "the most popular browser in the world". Another user stepped in to defend him, saying that if the site changed and the browser didn't change, then clearly it was the site to blame.

This is where immediate and underlying causes come in.

The immediate cause of the problem, in this case, was that the site changed to a different authentication system.

The underlying cause was that the browser wasn't up to the task of dealing with the different authentication system.

Thus, while changing the website back would have the effect of the browser understanding the authentication system again, it would not solve the problem of the browser not understanding newer authentication systems. It would be like replacing the lawn with pliable rubber - it would reduce the risk of getting a broken arm from falling on that lawn, but it wouldn't reduce the risk of falling off the ladder in other areas.

The moral of the story: the apparent cause isn't always the important cause or the right problem to fix.