Huffman Coding

Consider the following message: she sells seashells by the seashore

Determine the frequency distribution for each character in the message (including spaces). Just count and record the number of times each character occurs. Draw a Huffman tree built from the frequency distribution from the previous step. Demonstrate your understanding of Huffman coding by encoding the word seashells using your tree. In a Huffman tree was designed using frequencies taken from a given message, and for every new message we would have to build a new tree. Imagine instead that we had built one tree to encode all different messages (in English, say). How would you determine what that Huffman tree should look like? What would be the advantages and disadvantages of building one tree to encode all messages?