According to early reports, Monday's North Korea event certainly seems like a deliberate explosion in the right place. However, it was too small to be a successful Hiroshima-class crude explosive device, by a factor of three or four. The reported estimates of Richter magnitude spread from 4.5-5, and the standard conversions to explosive yield suggest a yield of 2-6 kiloton-equivalents of TNT. Most of the latest Richter magnitude estimates have come in the low half of the 4.5-5 range, so it seems likely that the yield was 4 kilotons or smaller.
That's a lot of energy, much larger than the 2006 North Korean test, but it still falls far short of an expected 12-20 kiloton yield of a crude Hiroshima-style device. For comparison's sake, the first nuclear tests of all other nations that are self-announced members of the nuclear club had larger yields than this latest North Korean test.
Because the expected Hiroshima-style explosion didn't occur, there are four options as to what did happen during the test:
The first option is the most likely case given what is publicly known about North Korean diplomacy and technology.
It's important to realize that nuclear tests in the past 15 years have primarily been demonstrations of power and national will, rather than driven by engineering. The two different kinds of simple bombs are uranium 235-based and plutonium-based. The United States used one bomb of each type on Japan in the closing months of World War II, with roughly similar explosive yields. The technology for a uranium 235 bomb is so simple, relatively speaking, that such a device need not be tested. To wit, the United States dropped the uranium 235 Hiroshima bomb without testing its design in a controlled detonation. Conversely, building a plutonium-based bomb represents a technical challenge because the critical mass can blow apart in a split second before the detonation reaches max efficiency. Once U.S. weapon designers confirmed the trickier detonation scheme of a plutonium device in the famous Trinity Test in 1945, confidence in the simpler uranium 235 detonation design seemed justified.
The technology for making the raw material for each device differs. Uranium 235 must be separated from the heavier isotope uranium 238 to make a bomb, so a country that buys or builds high-tech centrifuges is investing in those types of bombs. The term "enriched uranium" refers to uranium metal with an increase of the 235 isotope that's sufficient to allow an explosive chain reaction to occur. (Complete separation of the two isotopes is neither feasible nor necessary to create a weapon.) Plutonium comes from nuclear reactor waste, many stockpiles of which North Korea possesses. In other words, North Korea is attempting to make the trickier of the two devices. When arms control experts estimate North Korea to have a nuclear stockpile of 6-8 weapons, they are converting the likely amount of reactor waste that the country has produced into plutonium bombs of the World War II-era design.
When the device is built correctly (the earliest technology used a sphere with precisely timed explosions to press an array of small plutonium masses into one critical mass), the yield is 10-20 kilotons. It's unlikely that North Korea has skipped this step and gone on to more sophisticated designs that involve other elements (for instance, tritium) to generate a smaller yield from a smaller plutonium mass. China didn't. Its tests grew in size and sophistication in the 1980s and 1990s until Beijing induced seismic events that registered higher than 6 on the Richter scale. Then they stopped. Thus, I would rule out option number two.
Option number three also seems unlikely. The 2006 test was never proven to be fake, and more largely, there's no reason for Pyonyang to fake a test if it could at least attempt a real nuclear detonation. Nor has any world leader ever publicly called out North Korea for executing a failed or fake test in 2006; such a response probably would have pushed Pyongyang to attempt a second test much more quickly. It's the same reason why the United States and Europe--despite seismic data to the contrary--didn't call India's thermonuclear bluff in 1998; they wanted to reduce tensions, not raise them.
The last option--detonation in a cavity--makes no sense at all. The common assumption behind the Comprehensive Test Ban Treaty is that countries would attempt to evade detection. But in the twenty-first century, what good is a nuclear weapon if it can't be used to openly check adversaries? Plus, a country doesn't even need a military useful device to do so. With cable news, the ability to detonate something fissionable does the trick--at least in a political sense.
So my guess is that North Korea tried and failed to get a simple plutonium bomb to detonate correctly. Make no mistake--an inefficient nuclear weapon is nothing to dismiss. Even at the low end of its estimated yield (2 kilotons), the May 25 test released as much or more explosive energy than the largest conventional-explosive air raids during World War II. But one should be mindful of the technical challenges North Korea still faces in carrying out the threats implied by its deliberate pairing of its explosive test with test missile launches.