No label defined (Q4379977)

This is an interesting survey talk on the problem of classification of algebraic surfaces of general type. It is well known that both numerical invariants \(K_S^2\) and \(\chi(\mathcal O_S)\) of a minimal surface \(S\) of general type satisfy \(K_S^2\geq 1\), \(\chi(\mathcal O_S)\geq 1\) and \(K_S^2\leq 9\chi(\mathcal O_S)\). When \(\chi(\mathcal O_S)\) is sufficiently large, it is shown that there exist minimal surfaces \(S\) of general type whose numerical invariants can fill the preceding region except near the upper bound \(K^2<9\chi\). But for smaller \(\chi\), the problem is more complicated. On the other hand, since the canonical model of a surface is an important birational invariant, the problem of finding canonical surfaces becomes attractive. In this paper, the author raises two questions: 1. For which \((x,y)\) in the region described above is there a surface with an almost generic canonical projection (equivalently, when is the canonical system without base points and yielding a birational map)? 2. When is there a surface \(S\) which is embedded by a canonical system? In fact, the recent advances in homological algebra can be applied to this purpose. This paper starts on basics of homological algebra of polynomial rings. Then the canonical surfaces in low dimensional projective spaces are discussed in detail. Many interesting examples with small numerical invariants are presented. In the last section, applications of homological algebra to the effective construction of algebraic varieties with given invariants are discussed. For the entire collection see [Zbl 0882.00032].