Nuclear engineer here. I was a reactor engineer at Susquehanna duriing the construction phase. On the day of the accident, I was part of a team of nuclear fuel and reactor engineers from Susquehanna/PP&L visiting Carolina Power & LIght in Raleigh, NC and Southern Company in Birmingham, AL to discuss plant process computers and nuclear fuel when the accident occurred. Information was sketchy until we got back to PA.
On the way back, I stopped in to see my grandmother who lived in Mechanicsburg - about 15 air miles from TMI. She asked what she should do. I told her to just stay inside for a few days. Very little radiation escaped from the containment or other facilities, so unless things went really bad, there was very little danger to the public.
All the long term health effects studies have shown that there was no increase in serious disease in the 30 years following the accident.
The reactor vessel held the melted fuel (just like it was supposed to do in an accident scenario) and the containment held the airborne/waterborne radiation. There were enough redundant safety systems in place so that despite screw ups by the operators and management, the threat to the public was minimal.
I've read several reports on the TMI accident in the last 30 years and they seem to come to different conclusions about root causes. It seems that the root cause is a combination of reactor design (Babcock & Wilcox designed it), training, balance of plant design, inexperienced management, inability to see what the issues were, and regulatory.
Here's one cause that is rarely discussed: GPU/Met Ed were in a huge hurry to get TMI-2 on line by the end of 1978 so they could get it into the rate base in 1978 and collect revenues based on it being in the rate base. They probably rushed it and paid the price in March of 1979. They declared TMI-2 to be in commercial operation on December 31, 1978.
For those who want to build more nukes, well, you can do it in China. China put four Westinghouse AP1000 units on line in the last several years with an average construction time of about 9 years - even with a slowdown due to the Fukushima (Japan) incident.
On the other hand, I worked on the AP1000 units being built in Georgia and South Carolina. We started work on these units in 2008. The first unit in Georgia might come on line in 2023 (maybe). The second unit may come on line 2024. The two units in South Carolina were halted when the owner (South Carolina Electric and Gas) filed for bankruptcy because they ran out of money to fund the project.
One other thing to consider is that much of the world's uranium is controlled by Russia.