Computation of topological invariants for real projective surfaces with isolated singularities (Q480681)

The authors extend their prior work (see, e.g., \textit{E. Fortuna} et al. [Appl. Algebra Eng. Commun. Comput. 16, No. 5, 271--292 (2005; Zbl 1094.14050)]) to real algebraic surfaces with isolated real singularities. (There are no constraints on the complex singular locus.) Given such a surface \(S\subset\mathbb{P}_{\mathbb R}^3\), denote by \(T\subset\mathbb{P}_{\mathbb R}^3\) the result of ``embedded topological desingularization'' of~\(S\), i.e., the nonsingular (topological) surface obtained by spreading apart the topological local branches at each singular point. (Clearly, the abstract singular surface \(S\) can be regarded, not quite canonically, as a quotient of~\(T\) obtained by contracting finitely many finite sets.) Then, the following topological invariants of the pair \((\mathbb{P}_{\mathbb R}^3,S)\) are computed in terms of the defining equation of~\(S\) (assuming that the latter is square free and defined over \(\mathbb Q\) or, more generally, over a computable subfield of \(\mathbb R\)): {\parindent=6mm\begin{itemize}\item[1.] the set of connected components of~\(T\) and their Euler characteristics, \item[2.] the quotient projection \(T\to S\), \item[3.] the set of components of \(\mathbb{P}_{\mathbb R}^3\setminus T\) and their adjacency graph, including the component adjacent to the one-sided component of~\(T\) (if present), \item[4.] for each component of \(\mathbb{P}_{\mathbb R}^3\setminus T\), whether it is contractible in~\(\mathbb{P}_{\mathbb R}^3\), \item[5.] the number of isolated singular points of~\(S\) in each component of \(\mathbb{P}_{\mathbb R}^3\setminus T\). \end{itemize}} Note that items 3. and 4. above suffice to detect the contractible components of~\(T\); it appears that the contractability of those of~\(S\) is not studied.