At this point, it looks like there will hopefully be 3 slice sampling variants implemented in dynesty. The first is the multivariate variant currently implemented where we execute a Gibbs-style update along the principle axes of the bounding ellipsoid. The second is a random variant similar to PolyChord where we slice along a random axis drawn from the ellipsoid. The third is the Hamiltonian variant where we slice in time along the entire trajectory. These latter two should be given different names (say rslice, and hslice) to accompany their implementation.
At this point, it looks like there will hopefully be 3 slice sampling variants implemented in dynesty. The first is the multivariate variant currently implemented where we execute a Gibbs-style update along the principle axes of the bounding ellipsoid. The second is a random variant similar to PolyChord where we slice along a random axis drawn from the ellipsoid. The third is the Hamiltonian variant where we slice in time along the entire trajectory. These latter two should be given different names (say rslice, and hslice) to accompany their implementation.