Sampling problems for randomly broken sticks