Not entirely accurate. We know variolation will provide immunity. The main risk is to the volunteer, not to others. A vaccine might not provide immunity. That would make the patient a vector for the disease and endanger others.
> Do we really? I don't keep up with the news, did we already dismiss those reports about re-infection? Does it last long enough to be globally useful?
It seems to be extremely rare, and it's hard to know how many of those cases are due to incorrect initial diagnosis or people with odd immune systems. Immunity seems to be much greater than that conveyed by vaccines (which protect 85-95% of recipients).
> The main risk, sure, but the volunteer will become infectious. Vaccines constrain the risk to volunteers a lot better.
In vaccine challenge trials, people are exposed to the virus and kept quarantined until after the incubation period. Those who aren't protected by the vaccine must remain quarantined until their immune system defeats the disease. It would be the same for deliberate infection. Volunteers wouldn't be allowed to leave until the virus is no longer detectable. Hanson makes this clear in his blog post (linked to at the top of this thread).
That said, most vaccine trials are not challenge trials. Researchers give patients the vaccine and wait a while to see how many of them naturally contract the disease. During that time, the patients may or may not have immunity and can potentially infect others.