Silly example. It compresses the phrase "do or do not" that has 6 symbols (d, o, r, n, t, space), builds a huffman tree just for these and then assumes significant compression by using 8 bits per character for the uncompressed case.
I think it still suffices as an example because it would be easy to infer how this scales up to an entire book. Finer details are left out, but is that the sort of detail that should be present in a very short introductory article?
My intention was to pick an example that produced a small tree with only a few leaf nodes (so that the diagram was easy to follow), but still contained some duplication.
My hope was that somebody new to the concept could then infer the results for larger inputs.
I did not intend to imply that this would be a valid use case for building a Huffman tree in practice.