Why there is no ``Fundamental theorem of calculus'' for the Riemann integral (Q686409): Difference between revisions

Let \(\mathcal A\) be a linear space whose elements are continuous real-valued functions on \([a,b]\subset\mathbb{R}\) and such that \(\text{id}_{[a,b]}\in{\mathcal A}\), \({\mathcal N}\neq\emptyset\) a collection of subsets of \(\mathbb{R}\) such that \(\text{int } N=\emptyset\) for \(N\in{\mathcal N}\) and \(M\cup N\in {\mathcal N}\) for \(M,N\in{\mathcal N}\). Suppose that \(F\in{\mathcal A}\) and \(N\in {\mathcal N}\), \(N\cap (a,b)\neq\emptyset\) imply \(F(N\cap (a,b))\in{\mathcal N}\). A function \(f: [a,b]\to\mathbb{R}\) is said to be \(({\mathcal A},{\mathcal N})\)-integrable iff there exist \(F\in {\mathcal A}\) and \(N\in{\mathcal N}\) such that \(f(x)\) is the right derivative of \(F\) at \(x\) whenever \(x\in (a,b)-N\); then \(F(b)-F(a)\) is independent of \(F\) and it is said to be the \(({\mathcal A},{\mathcal N})\)-integral of \(f\). The author shows that there are no \(\mathcal A\) and \(\mathcal N\) such that the class of \(({\mathcal A},{\mathcal N})\)-integrable functions coincides with that of the Riemann integrable functions or that of the regulated functions in the sense of Bourbaki. [Theorem 1 is a special case of (7.1) in \textit{S. Saks}: ``Theory of the integral'' (1937; Zbl 0017.30004)].