This is careful work. And a lot of it. We need to account for several KINDS of infinite sequences. Honestly, we will be duplicating a lot of the functionality of UAdic/IAdic/RAdic/ZAdic. We WANT the ideas of finite/signed-finite/repeating/uncertain, respectively. Think of those structs as functionality we want to take, as we go from finite to infinite Sequences. We don't want to take the arithmetic, the ops, of those structs. We want to take the digit manipulation from them: splitting Sequences and indexing and such. Look particularly to the HasDigits trait and see if we can generalize or split that functionality and take it to the Sequence side. However, we might not need to re-implement ALL of the nitty gritty of these structs either. Because we can do interesting mapping and zipping of Sequences, we could e.g. make periodic Sequences by special combinations of mapping and zipping constant Sequences. Not that we NEED to do things like this, just that it's possible. Interested to see where this goes!