• discrete distribution -- uniform distribution has maximum entropy
  

• continuous distribution -- Gaussian distribution has maximum entropy for a given mean and variance

• convex shape of the entropy functions make it easy for optimizations