Methods of comparison of families of real functions in porosity terms (Q2288419)

Given a real interval \(I\), let \(\mathcal{X}\) be defined as the space \[ \mathcal{X}:=\{f\colon I \to\mathbb{R}\} \] equipped with the metric \[ \rho(f,g) := \min\{1,\sup\{|f(x)-g(x)|\colon x\in I\} \] of uniform convergence. Roughly speaking, the concept of a porous set is a stronger version of the concept of a nowhere dense set with more control on the size of the ``holes'' in the set. There is a number of variants of porosity which vary in their strength including lower and upper porosity. These concepts are discussed in detail e.g. in [\textit{L. Zajíček}, Abstr. Appl. Anal. 2005, No. 5, 509--534 (2005; Zbl 1098.28003)]. In this article the authors compare various families of functions in \(\mathcal{X}\) by checking whether one of them is porous in the other one. These questions complement the ones considered in [\textit{F. Strobin}, J. Nonlinear Convex Anal. 13, No. 2, 351--361 (2012; Zbl 1246.47014)] where similar questions were investigated for certain classes of continuous functions in Hilbert spaces. One of the main results of this paper is a sufficient condition for strong lower porosity of one family of functions in another one which is given in terms of existence of a certain function in the larger family. This result is then applied to compare various families of real functions including continuous, absolutely continuous and Baire one functions. The second part of this paper is concerned with families of path continuous functions. Also in this context a sufficient condition for strong lower porosity of one family of functions in another one is given and applied to a number of examples.