Obviously a lot of smart and well-informed people have been thinking about this. Many, like Juan Cole, think that the Iranians are nowhere close to a bomb; ThinkProgress is slightly less sanguine. They are taking the trouble to make this argument because the US is claiming that it would only take 16 days for Iran to make a bomb. There are all sorts of reasons to disbelieve this particular claim: a history of crying wolf, an apparent misunderstanding of the concept of significant figures... Still, is it more like ten days, or ten years? Steinn Sigurðsson looks at the problem as a physicist, and isn't optimistic.
I don't know Iran; I don't have access to any classified information on nuclear weapons. I do know something about physics... First of all, Iran is clearly been working on putting together a full nuclear cycle for about 20 years That means they want to be able to do it all in-house: mining, enrichment, burning, plutonium extraction, power generation and bomb production. It is clear that they did the science in the early-to-mid-90s, they tested centrifuges, built small high neutron flux reactors and got small amounts of plutonium extracted. So, they learned Pu chemistry, what isotopes you get with different burns, and maybe some metallurgy. They then set up centrifuge halls and played with an AVLIS (laser isotope separator). They also ordered a 1GW reactor from the russians, and refined uranium oxide (aka "yellowcake") into both uranium tetrafluoride, uranium hexafluoride and uranium metal. Supposedly several tons of uranium oxide were processed. Now: there are two ways to make bombs, at the basic level. Get highly refined uranium-235 metal; or, fairly pure plutonium-239. In kilogram quantities. U-235 bombs are simple and need not be tested. "A grad student could make one of those". Pu-239 bombs are notoriously fickle and are said to need testing (although maybe not so much any more...)
Read the whole thing. Hofstadter's Law says "It always takes longer than you think, even when taking into account Hofstadter's Law." For nuclear weapons, unfortunately, the word "longer" should be replaced by "shorter." Historically, we always underestimate the proximity of other nations to full nuclear capability (unless we're trying to cook up reasons to invade them). I don't know what to do about it, but there's every reason to believe that, left to its own devices, Iran will have some sort of bomb sooner rather than later.
