While that example partially demonstrates the techniques you'll need, it's not intended to be scalable.
If you want to split a string manually like this, you need to look for the next occurrence of the delimiter you're looking to split on, store the text before that location in a variable, store the position of the found delimiter, look for the next delimiter position, store the text between those delimiters to a variable, store the position of the found delimiter, and so on.
In what context are you looking to split this text? If it's for a command phrase you're speaking, you could use the "{CMDSEGMENT:}" token instead.