Hi Ralf, your point is the exact reason why the binomial random variable is not suitable to model the number of clapping. Binomial r.v. can only take either 0 or 1, however, as you said, we will have 0 to n readers per hour.
Your second question: If we model the success probability by hour (0.1 people/hr) using the binomial andom variable, the model output will be like, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0.
(9 zeros and 1 one.
0.1 people per hour means — there will be only 1 person for the 10-hour period.)