On the two-dimensional pointwise dyadic calculus (Q1385381)

\textit{F. Móricz}, \textit{F. Schipp} and \textit{W. R. Wade} [Trans. Am. Math. Soc. 329, No. 1, 131-140 (1992; Zbl 0795.42016)] proved that if \(f\) is integrable on the unit square \({\mathbf Q}\), then the Cesàro means \(\sigma_{2^n,2^m}f\) of the double Walsh-Fourier series of \(f\) converge to \(f\) almost everywhere on \({\mathbf Q}\) provided \((n,m)\to\infty\) in some positive logarithmic cone, i.e., there is a \(\beta>0\) such that \(| m-n|<\beta\). There is an analogy between the fundamental theorem of dyadic calculus and summability theorems for Walsh-Fourier series discovered by \textit{F. Schipp} and \textit{W. R. Wade} [Appl. Anal. 34, No. 3/4, 203-218 (1989; Zbl 0727.42020)], when they proved a fundamental theorem for dyadic calculus for all \(L\log^+L({\mathbf Q})\) functions whose double Walsh-Fourier coefficients vanish on either axis. Here, the author extends this fundamental theorem of dyadic calculus from \(L\log^+L({\mathbf Q})\) to \(L^1({\mathbf Q})\) by restricting the way \((n, m)\to\infty\). Namely, he proves that the limit of the dyadic partial derivative operators of the indefinite dyadic integral of any \(f\in L^1({\mathbf Q})\), as \((n, m)\to\infty\) through some positive logarithmic cone, equals \(f\) almost everywhere on \({\mathbf Q}\), provided the double Walsh-Fourier coefficients of \(f\) vanish on either axis. The proof, which requires some very tenacious estimates, involves showing that the corresponding maximal operator is of weak type \((1,1)\).