Inverse real closed spaces (Q1182728)

The author has built up to a theory of ``real closed spaces'' [Mem. Am. Math. Soc. 397 (1989; Zbl 0697.14015)]. These are obtained by glueing affine pieces, which are proconstructible subsets of real spectra, equipped with their sheaves of abstract semialgebraic functions. They may be seen as an abstraction of the ``locally semialgebraic spaces'' of \textit{H. Delfs} and \textit{M. Knebusch} [see Lect. Notes Math. 1173 (1985; Zbl 0582.14006)]. \textit{M. Knebusch} was led to introduce also the ``weakly semialgebraic (w.s.a.) spaces'' [see Lect. Notes Math. 1367 (1989; Zbl 0681.14008)], in order to have homotopy theory working nicely. The ``inverse real closed (i.r.c.) spaces'' may be seen as an abstraction of these w.s.a. spaces, and they throw new light on these; the relationship between the abstract and geometric spaces is the main theme of the paper. For i.r.c. spaces, the affine pieces have to be glued along closed subspaces; this is done by considering on these affine pieces the inverse topology, where the closed constructible subsets form a basis of open sets (this explains the name). The author studies general properties of the category of i.r.c. spaces, and discusses notions of quasi- compactness, quasi-separatedness, finiteness \dots The formal properties are very much like those for schemes or for real closed spaces, but the geometric content differ. For instance one main result is that any quasi-separated quasi-compact i.r.c. space is affine.