Wondering why long vowels are not treated as simply two letters, for example. So
Actually, they are considered to be two vowels. This is discussed in The Representation of Vowel Length, but the really short answer is that long vowels are typically considered to be a single segment with two "timing units", hence the two-ness. There is a rich literature noting that sometimes, "long" segments behave like there are two of them, and sometimes they act like there is one of them. There has been a bit of a puzzle trying to figure out exactly what processes will treat a long thing like one thing, and what will treat a long thing like one thing. For instance, a rule that rounds a vowel next to a labial consonant (such rules exist) will round the entirety of a long vowel. Charles Pyle and Michael Kenstowicz wrote on this topic in the early 70's, and Pyle wrote his dissertation on the topic.
The view that has been pretty much universally accepted since 1978 is that a long segment is usually a single segment with special suprasegmental structure (most usually, these days, "a mora", sometimes "a skeletal slot". However, there are languages that actually contrast "long vowel" from "identical vowel cluster", so there is a difference between [o:] and [oo], where [o:] is a single syllable long vowel, and [oo], better notated as [o.o], is a two-syllable sequence.