The 128 ASCII characters can be represented with 7 bits rather than 8

[avalloud17]

Is this as you originally intended?

[klevasseur]

I would presume that there is a parity bit added. I didn’t get into this issue at this point in the book, but it’s hinted at by part (c) of the problem.

[avalloud17]

There is no such thing as a universal standard for character encoding. On some platforms that will be some Latin codepage, on some others it will be UTF-8, on yet some others it will be UTF-16, and so on. Unless you specify the encoding, the exercise will be somewhat undefined. But adding the complexity seems gratuitous especially with the UTF encodings because now you have to make sure the combinations you choose to count in the graph are not escape codes that do not represent characters themselves, etc...

To avoid all that, another way would be to phrase the question in terms of "integers between 0 and 127, differing in two bits when written in base 2".