5.8 (Schleimer) — Why SnapPy works in practice
“Give a rigorous explanation for why SnapPy works so well in practice.” The question asks for an explanation, not a specified theorem about a particular class of inputs; it does not define a success criterion or quantify over manifolds, triangulations, or computations. Problems in Groups, Geometry, and Three-Manifolds
Background
A central practical advantage is that hyperbolic geometry can often be encoded by an ideal triangulation and tetrahedron shape parameters. The gluing and completeness equations for a triangulation are polynomial equations in those parameters; a solution with all tetrahedra positively oriented describes a geometric triangulation. Such triangulations make a manifold’s geometry accessible to computation, and are central to SnapPy’s calculations of hyperbolic structures and their geometric verification. Infinitely many virtual geometric triangulations
The computational pipeline is deliberately not a single all-or-nothing numerical test. SnapPea’s approximate floating-point solutions can identify promising triangulations, and randomization by local moves such as Pachner moves can produce alternative triangulations when the initial one is not geometric. In the census verification procedure, failure to find a candidate after a bounded number of attempts is explicitly inconclusive, not evidence of non-hyperbolicity. When a candidate is found, rigorous interval methods can certify shape parameters and geometry. Thus numerical search helps locate a certificate; it is not itself the certificate. The cusped hyperbolic census is complete Random veering triangulations are not geometric
There is also a concrete history of increasingly extensive census computations: the cusped census began with manifolds built from at most five ideal tetrahedra and was later expanded through eight. This provides a setting in which practical computational success can be tested against exhaustive mathematical claims, rather than judged only by examples found by the software. The cusped hyperbolic census is complete
Progress summary
Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.
Solutions 0
No solutions have been posted yet.