On modular homology of simplicial complexes: Shellability (Q5930027)

Given a simplicial complex \(\Delta\) and a coefficient domain \(F\); if \(F\Delta\) is the (free) \(F\)-module with basis \(\Delta\), then the inclusion map \(\partial: \tau\mapsto \sigma_1+\cdots+\sigma_k\), with \(\sigma_i\) a codimension-1 face of \(\tau\in\Delta\), is not a boundary operator but, e.g., if \(F=Z/(p)\), then \(\partial^p= 0\) nevertheless and one can consider a homology-like theory, here referred to as modular homology, a subject of interest to these authors for some time already in previous publications, where the simplical complex \(\Delta\) was often of the form \(\Delta(P)\), the order complex of a finite poset \(P\) whose simplexes are the maximal chains with faces chains of arbitrary length, the posets \(P\) being of some prescribed type, e.g., \(P\) is a Boolean algebra. In the current paper the issue is taken up again and the condition of \(P\) is that \(P\) be shellable, leading to a considerable amount of detailed information of a general kind, e.g., every shellable complex with \(h\)-vector \((h_0,\dots, h_k, 0,\dots, 0)\) has some exact \(p\)-modular homological sequence if \(k\leq p\), or more particularly, the modular homology for complexes \(\Sigma^{n-1}_k\), even more particularly for \(p=3\), \(n=8\) as \(1\leq k\leq 8\) as in table I. These results can be translated into combinatorial information, such as \(f\)-vectors etc. Also, each step on this road leads forward while suggesting other steps yet to be taken, such as the modular homology of finite buildings promised next.