@ Dekker, what use does tritium have in optics???
@ Androgeos Exeunt, fusion should be relatively safe in the sense that a run-away reaction is exceedingly unlikely. Virtually every fusion reactor that works even a little relies on "pulsed" fusion. Reaction duration is usually measured in microseconds. It takes extraordinary effort to even get the reaction to last that long. I took a course on plasma engineering about 13 years ago, and the basic problem with fusion reactions is that the closer you get to conditions where a fusion reaction is "sustainable," the more efficient all of the various mechanisms for cooling the plasma become.
As for radiation, charged radiation (alpha and beta) are fairly easily dealt with. Your main problem is neutron radiation. It is eventually going to make your reactor materials radioactive as all get out. It is possible to shield against neutron radiation, but not for the materials that are part of the reactor itself. That results in a kind of "corrosion" as materials in the reactor transmute / decay into other elements. I suppose this isn't a new problem. Fission reactors have to deal with this as well, but they usually are filled with a neutron moderating fluid. That won't be possible in a fusion reactor.
It may be possible to make a useful fusion reactor (in the sense of power generation) at some point, but I wonder if it is going to be more trouble than it is worth. At least planet-side. The only place where I can see it really being worth the extreme pain is in space, where the mass of your fuel becomes such a bottleneck to operational range.