You really have to put things in context, otherwise you get into the "think of the children" fallacy, where everything is worth the effort if it saves one life.

We die of three causes: cancer, heart problems and strokes. Well over 30% die of cancer. So this means some of those 200 will die of cancer before they die of something else (some of them will recover and live to die from heart attack later). This makes the new deaths a really tiny percentage, which means policy-wise it's hardly worth discussing.

I'm certainly not trying to imply any policy prescriptions. I'm strongly in favor of nuclear power. But if there's risk we should know what it is so it can be distributed/compensated for.

I'm not saying it's not worth the cost (or even trying to discuss things in those terms), but I'm not convinced this outcome is insignificant. If this article is right, many of these people will die N years sooner than they would have otherwise. How high does N have to be before it gets worth discussing?

The risk for the GE MkI was (is) well known, but physicist-socipath Seaborg chose to hand out the operating license:

http://news.ycombinator.com/item?id=4401744

It is pretty bad, but people have seriously argued that the evacuation itself killed more people: Japan (like any other highly populated First World country!) is stuffed full of old or sick people, and when you force them to evacuate quickly and be separated from all their stuff, their doctors and medical records, their local pharmacies, and subject them to a ton of stress in every respect - well, quite a few people are going to kick the bucket.

I think the evidence on that point is incontrovertible. On the other hand it would have been politically impossible to refuse to evacuate people in the area.

Over 15K people were killed in the quake and tsunami. On the scale of "badness", ~200 projected cases is not a scale tipper. Not to trivialize the additional projected cancer risks, just to put it in perspective.

what other industry gets to benchmark social license against this tsunami? What other unacceptable risks should be weighed up against 15,000 deaths "to put it in perspective"? On the scale of badness, how many deaths does an industry need to be responsible for until its "a scale tipper"? This is a very poor way of weighing up the very real ongoing risks and hazards presented by the nuclear power industry.

My main point in putting it "in perspective" was not to dismiss the statistically projected increase in cancer rates, but to show that if there is anything to panic about first in Japan, it is the systemic failures that may have lead to a higher death toll in the earthquake and tsunami, such as tsunami walls of insufficient height, and "safe zones" that were not safe enough for the surge height. The surge estimation, modeling, contingency and responsiveness "faults" for the general population were of the same kind that affected the unanticipated failure modes of the power plant.

The plant did not fail because "nuclear was bad", it failed because it wasn't designed to be hit by a tsunami of that magnitude; which if the same reason for many of the direct tsunami-related deaths. With that in mind, Japan needs to get better at tsunami planning and engineering, like they generally have with direct earthquake planning and engineering.

Now, as to comparing energy producing industries, the mortality metric would be deaths per kilowatt hour; in which nuclear fares extremely well compared to mainstays like coal, even if 200 additional cancer deaths out of a "normal" 4400 for the given population are added in.

Other industries would of course be considered on merit. I don't know if any so thoroughly under attack from people with a poor grasp of the topic (or supported by people with such a poor grasp of the topic...)

What other industries would it be relevant to benchmark like this?

Why is it a poor way of weighing up the hazards?

Every energy source kills people, even solar and wind kill people. This guy did a collection of statistics (I think they are each by a different body so they're not directly comparable but should still suffice for orders-of-magnitude.) http://nextbigfuture.com/2011/03/deaths-per-twh-by-energy-so...