Density-zero conjecture for local defect sets
Establish that the local defect set has relative density zero for every visibility datum whose branch-locus counting function has polynomial growth, whose local visibility conditions each possess a relative density, and whose large-prime tail satisfies a uniform estimate analogous to the paper’s large-prime-tail condition.
References
These observations naturally lead us to make the following conjecture.
Let be a visibility datum such that #(;X) has polynomial growth, every local visibility condition has a relative density, and the corresponding large-prime tail satisfies a uniform estimate analogous to eq:large-prime-tail. Then $\mathfrak{d}{bad}()=0$.
eq:large-prime-tail:
The examples in Section \ref{sec:sparse} suggest the following conjecture.
Let n\geq 2, and let be a visibility datum whose branches are defined for all t>0 and have the form
\phi_\alpha(t)
\bigl(a_1(\alpha)t{w_1},\dots, a_n(\alpha)t{w_n}\bigr),
where w_1,\dots,w_n\in_{\geq 1} and a_i(\alpha)\in_{>0}. Assume that $\gcd(w_1,\dots,w_n)=1$ and suppose that () is sparse. Then, for every prime p, the relative density \mathfrak d_p{\operatorname{vis}() should exist, the Euler product of the local densities should converge, and the relative density of globally visible points should exist and satisfy
\mathfrak d{\operatorname{vis}()
\prod_p \mathfrak d_p{\operatorname{vis}(), where both the global density and the local densities are taken relative to ().