I'd been expecting this question, although that doesn't mean we're prepared for it. The Dept. of course has no official position, and Lee and I are not particularly experts. We especially have no particular ideas about how to deal with the current crisis. What we say below should only be taken as a first rough pass, just to get you started on some of the issues.
On your second question, even non-experts have heard of the fundamental problem with this type of reactor design. It requires almost continuous active cooling to avoid a disaster. As we're seeing (and not to much surprise) when things go wrong they often do so in ways the cripple many systems at the same time. Reactors should be designed with more passive
"fail-safe" protection (http://en.wikipedia.org/wiki/Passive_nuclear_safety
) to reduce the chance that they'll run into disasters if the ability to actively intervene is lost. (That doesn't mean that no disaster could conceivably happen, just that the vast majority of obvious things that could go wrong wouldn't be catastrophic.)
The CANDU reactors (http://en.wikipedia.org/wiki/CANDU_reactor
) are said to be examples of ones that are a little bit less prone to catastrophic failures. The deuterated water coolant plays a larger role in sustaining the chain reaction than does ordinary water, so that loss of coolant helps shut the reactions down more than in the plain-water systems. It's also claimed that the sorts of geometries that the fuel rods are likely to form in a breakdown do not sustain chain reactions.
There are newer designs which are believed to be more fail-safe than any of the boiling water or pressurized-water designs. (http://en.wikipedia.org/wiki/Pebble_bed_reactor#Criticisms_of_the_reactor_design
) It's a shame that the nuclear industry haphazardly built up old designs rather than slowly ramping up toward standardized, reliable and relatively safe newer designs via a systematic research program.
Of course, it would be foolish to claim that any design is perfectly safe. Airplanes can crash into reactors, people can hack computer systems, blow up containers, etc.
On particulars- obviously reactors shouldn't be on known seismic fault zones. Ones near seas should have extra sea walls, well beyond what can be practically done to protect larger areas. The big issue, however, remains active vs. passive protections.
All of these obvious precautions have been routinely violated. The reactor upwind of our town is basically the same design as the ones that are now melting down, although it has a larger inner containment vessel. It's in no danger of a tsunami, but it could possibly face a major earthquake.
(published on 03/16/2011)