Yeah, I think so too. I was originally just describing the way I saw things, then also added my own terminology, then saw the two weren't exactly the same.
I don't really like the terms "Teleological vs. Ontogenetic", because:
- nobody can remember those pompous names
- Splitting algorithms into "those who imitate nature's process" and "those who try to get the same result through another process" is a bit narrow, it assumes that you want to imitate nature in the first place, which is often wrong (I don't want to imitate nature, I want to make a fun game!)
… but then, the definition of what's "top down" and what's "bottom up" is a bit unclear too, not to mention that those two terms may be a bit emotionally loaded (which is still better than "incomprehensible").
Maybe a more fruitful discussion would be whether a characteristic is a provable, explicit outcome of the algorithm, or whether it's "just" a frequent side-effect. For example, I don't know whether Pelin Noise is "top-down" or "bottom up", but I know it doesn't explicitly try to get a certain topology.
