There are always caveats with these types of indexing methods, especially if you require dynamic high-performance indexing and support for rectangles/polygons. You can only move the edge cases around, there is no way to eliminate them. This allows you to tailor an indexing scheme for specific workload assumptions but this obviously breaks down if you need an algorithm that generalizes to many workloads and data models. There isn't just one type of edge case with this type of indexing, there are several which may or may not be relevant depending on what you are trying to do.
Some research from the 1980s showed it is only possible to mitigate bounded categories of edge case, thereby improving generality, by indexing on complex higher-dimensionality embeddings. Mitigating more categories requires more dimensions and more complexity. However, no one could figure out how to construct these embeddings for even basic cases or deal with more practical curse of dimensionality issues, so that is largely forgotten (the researchers themselves made comments to the effect that they didn't think a tractable solution was possible). I've never even been able to find that literature in electronic form, unfortunately.
Like AI, it is an interesting open-ended problem space. You can prove that an elegant optimal solution is not tractable so it ends up being a search for asymptotically optimal algorithms that become exponentially more complex the closer you get to optimal.
Some research from the 1980s showed it is only possible to mitigate bounded categories of edge case, thereby improving generality, by indexing on complex higher-dimensionality embeddings. Mitigating more categories requires more dimensions and more complexity. However, no one could figure out how to construct these embeddings for even basic cases or deal with more practical curse of dimensionality issues, so that is largely forgotten (the researchers themselves made comments to the effect that they didn't think a tractable solution was possible). I've never even been able to find that literature in electronic form, unfortunately.
Like AI, it is an interesting open-ended problem space. You can prove that an elegant optimal solution is not tractable so it ends up being a search for asymptotically optimal algorithms that become exponentially more complex the closer you get to optimal.