By describing details of the design and operation of one nuclear power plant, Daniel Ford reveals a good deal about broader aspects of design, management, and regulation. Much of the book is surprising and even disturbing, and yet--to a practical mind at least--everything that Ford isolates and explains as problematic is, in principle, remediable. How? Precisely because factor X can be identified as having failed, it's possible to change factor X in order to keep it from failing in that way again.
So much for principle. As Ford makes clear, what went wrong at TMI could've been prevented ahead of time--it didn't have to happen. Yet it did. Though I'm no expert, I doubt that all the broader problems have been fixed after the fact either.
Given that the world currently gets a sizable proportion of its energy from nuclear power, it's basically impossible to turn off all the plants. Nothing can instantly replace all those gigawatt-hours of power-generating capacity. So something like TMI may happen again. Or something like the multiple partial meltdowns at the Fukushima Dai-ichi plant in Japan. (I have to sidestep Chernobyl, because of the special absurdities of its design and operation.) A good book-length study of Fukushima will probably be produced, if it hasn't already. In the meantime, Daniel Ford's report on TMI, though more than 30 years old already, is still a valuable study.