Why ? It looks like those ShakespeareIDs (SIDs) are much longer than UUIDs. So I am not sure how you reach the conclusion than the count of possible SIDs is sma

Why ? It looks like those ShakespeareIDs (SIDs) are much longer than UUIDs. So I am not sure how you reach the conclusion than the count of possible SIDs is smaller than UUIDs. The author mentions its a bijection, meaning for every UUID you can generate its corresponding SID and vice versa. This, in turn, means there is no higher probability of collision in the SID space than in the UUID one.