As I wrote before, I was worried about hurting my feet at first. After nearly six months, however, I have a better sense of the risks [2] and am much less worried. I've hurt my feet five times: four times from stepping on glass, once from pulling away a callous. The latter I've learned not to do: I leave my soles alone and they get sturdier on their own. The glass, though, has taken some getting used to. The first time I didn't recognize the feeling, and when sat down to look at my foot and figured it out I was kind of freaked out. I'd been walking on it for a while, and it was a little hard to get out. The more recent times, however, I'd learned to recognize the pain of glass shards. It is precise and localized, affecting only a tiny area of the foot and hurting only when weight is applied. Those three times I was able to quickly sit down with a tweezers (which I have in my pocket knife) and remove the glass. These were all tiny shards and came out easily.

So my experience with walking barefoot so far agrees with the evidence that it is healthier than walking in shoes.

[1] I would like to know how much of this is due to barefoot walking with the forefoot striking, how much to stretching, how much to exercising, and how much to "things just get better on their own". I don't know a good way to find out without risking major pain and not being able to walk or dance. But if, say, the exercises are not actually helping, I would like to stop doing them and regain that 10min/day.

[2] Several people pointed out that my previous risk calculation was wrong. I had been prompted to write by stepping on glass, which meant that my choice of time range was dependent on my observation. If I model this as "wait until you step on glass, then blog", then the right frequency to predict is 1/2 per time period, not 1 per time period.