Boundary

We say a boundary is some condition that splits an identifier into a list of words.

Some boundaries discard some substring at the location of an identified boundary, and others do not.

A boundary consists of three parts:

• the condition: a function that takes a slice of the identifier starting at location i to the end of the identifier, and returns it true if the boundary condition is met
• the index of the start of boundary start relative to the slice of the identifier
• the length of the boundary len

To identify the locations of an identifier that contain boundaries, we iterate through substrings of the identifier. For i from 0 to the length of identifier I, we create a slice S = I[i:]. If this slice meets the condition, we have identified a split in the identifier. The end of the previous word is at index i+start exclusive and the start of the next word is at index i+start+len inclusive.

Splitting

We say splitting is the process of converting an identifier into a list of words.

In pseudocode, you could implement splitting by a boundary b as follows.

function split(b: Boundary) {
    words = []
    last_word_start = 0

    for i in 0..n {
        S = I[i:]
        if b.condition(S) {
            last_word_end = i + b.start
            words.append( I[last_word_start:last_word_end] )
            last_word_start = i + b.start + b.len
        }
    }
    words.append(I[last_word_start:])

    return words
}

Delimiter Example

Suppose we wanted to split a snake case identifier into words. The boundary can be defined as follows:

• condition: is the first character in the slice equal to an underscore? S[0] == "_"
• start: 0
• len: 1

Examine the identifier I = last_byte_count. This boundary condition would be true at index 4. This means the start of the boundary would be at 4 + start, or 4. The length of the boundary is 1, so the end of the boundary is 4 + 1, or 5. Using these values to split the word, we would get I[:4] = last and I[5:] = byte_count. We would continue until we reach the end of the string.

Camel Case Example

Camel case is delimited by an empty string. We can define the boundary as follows:

• condition: is the first character lowercase and the second character uppercase? S[0].is_upper() and S[1].is_lower()
• start: 1
• len: 0

Because our condition checks the character before and after the split, we set a start of 1. And we don’t want to remove any characters while splitting, so the length is 0.