Interpolation and Sidon sets for compact groups (Q451981)

Since the seminal paper of \textit{W. Rudin} [J. Math. Mech. 9, 203--227 (1960; Zbl 0091.05802)], Sidon sets were the subject of intense research, one of the jewels of which was Drury's result on the union of two Sidon sets [\textit{S. W. Drury}, C. R. Acad. Sci., Paris, Sér. A 271, 162--163 (1970; Zbl 0194.42804)]. The book of \textit{J. M. Lopez} and \textit{K. A. Ross} [Sidon sets. New York: Marcel Dekker (1975; Zbl 0351.43008)] was the reference on this subject, but it appeared before the work of \textit{D. Rider} [Duke Math. J. 42, 759--764 (1975; Zbl 0345.43008)], and the papers of Pisier (1978--1983), and Bourgain (1983--1985), which drastically changed and improved our knowledge on Sidon sets. Until now, no other book was dedicated to that subject, except for the one by \textit{M. B. Marcus} and \textit{G. Pisier} [Random Fourier series with applications to harmonic analysis. New Jersey: Princeton University Press and University of Tokyo Press (1981; Zbl 0474.43004)], which is mainly on Gaussian processes and is rather difficult to read. The book under review at last fills a gap in the literature and gives a complete overview of the actual state of the art. Moreover, it concerns not only Sidon sets, but more generally interpolation sets (more specifically \(I_0\) sets), and it is the first book on this subject. A unifying thread of the book is the behaviour of interpolation sets with respect to the Bohr topology. This is motivated by the two main open questions on the subject: can a Sidon set be dense in the Bohr group, and is every Sidon set a finite union of \(I_0\) sets? (An affirmative answer to the second problem implying that Sidon sets are Bohr-closed in their underlying group). This book is very well written, with a pedagogic concern, and very pleasant to read. The chapters end with some exercises (I did not try to solve any!) and very interesting historical comments and indications on connected subjects. The first chapter is devoted to the classical Hadamard sets. Chapter 2 treats \(\varepsilon\)-Kronecker sets. A subset \(E\) of a discrete abelian group \(\Gamma\) is an \(\varepsilon\)-Kronecker set if, for every unimodular function \(\varphi\) defined on \(E\), there exists \(x \in G = \hat \Gamma\) such that \(|\phi (\gamma) - \gamma (x)| < \varepsilon\) for every \(\gamma \in E\). The main result of the chapter is the authors' result that \(\varepsilon\)-Kronecker sets are \(I_0\) sets if \(\varepsilon < \sqrt 2\); cf. [\textit{C. C. Graham} and \textit{K. E. Hare}, Math. Proc. Camb. Philos. Soc. 140, No. 3, 475--489 (2006; Zbl 1105.43003)]. In Chapter 3, the study of \(I_0\) sets begins. It is perhaps time to recall their definition. Note that one way to say that a subset \(E\) of a discrete abelian group \(\Gamma\) is a Sidon set is that every bounded function on \(E\) is the restriction of the Fourier transform of a measure on \(G = \hat \Gamma\). If the measure can be taken discrete, one says that \(E\) is an \(I_0\)-set. The two main results are \textit{N. J. Kalton}'s characterization [Colloq. Math. 64, No. 2, 233--244 (1993; Zbl 0838.43007)] and the Ryll-Nardzewski theorem: \(I_0\) sets are Bohr closed in \(\Gamma\). Chapters 4 and 5 study subclasses of \(I_0\) sets, namely \(RI_0\) sets (the interpolation measure can be chosen to be real), \(FZI_0\) sets (it can be chosen to be positive; \(FZ\) refers to the Fatou-Zygmund property), and their local counterparts \(I_0 (U)\), \(RI_0 (U)\) and \(FZI_0 (U)\) (the measure can be chosen to be supported by the Borel set \(U\)); these results are essentially due to the authors. The study of Sidon sets begins in Chapter 6, in which classical results are shown, as well as Bourgain's result that, for every positive integer \(k\), every Sidon set (which does not contain the unit element) is a finite union of \(k\)-independent sets, and Déchamps-Gondim-Bourgain's theorem that every Sidon set is a finite union of Sidon sets whose pace tends to infinity. Chapter 7 is devoted to Pisier's characterizations of Sidon sets [\textit{G. Pisier}, Adv. Math., Suppl. Stud. 7B, 685--726 (1981; Zbl 0468.43008); in: Topics in modern harmonic analysis, Proc. Semin., Torino and Milano 1982, Vol. II, 911--944 (1983; Zbl 0539.43004); Bull. Am. Math. Soc., New Ser. 8, 87--89 (1983; Zbl 0505.43002)], in particular, that a set (which does not contain the unit) is Sidon if and only if each finite subset of it contains a further subset of (constant) proportional size which is quasi-independent. The proof, with Bourgain's improvement, is that given in [\textit{J. Bourgain}, Ann. Inst. Fourier 35, No. 1, 137--148 (1985; Zbl 0578.43008)]. This easily implies Drury's union result and the solution of Rudin's problem that \(E\) is Sidon if (and only if) it is a \(\Lambda (p)\)-set with constant \(O\, (\sqrt p)\) for every \(p > 2\). The original proof of Pisier used Gaussian processes [\textit{G. Pisier}, Sémin. Géom. des Espaces de Banach, Éc. Polytech., Cent. Math., 1977--1978, Exposés No. 12, 13 (1978; Zbl 0388.43009)]. Chapter 8 concerns the size in the Bohr group of Sidon sets. A proof is given of \textit{L. T. Ramsey}'s result [Colloq. Math. 68, No. 2, 285--290 (1995; Zbl 0830.43014)] that, if there exists a Sidon set with a Bohr cluster point in \({\mathbb Z}\), then there is a Sidon set wich is dense in the Bohr group. The recent proof by \textit{J.-P. Kahane} and \textit{Y. Katznelson} [J. Anal. Math. 105, 363--378 (2008; Zbl 1153.43004)] of an old result of Katznelson is also given saying that, if one randomly chooses the integer \(n\) with probability \(p_n\), then the random set that one gets is almost surely a Sidon set with no Bohr cluster point if \(p_n = O\, (1/n)\), and is almost surely non Sidon and dense in the Bohr group if \(n p_n\) goes to infinity. In Chapter 9, the nature of the relationship between Sidon and \(I_0\) sets is explored. It is proved that Sidon sets are proportional \(I_0\) sets, and that they are proportional \(\varepsilon\)-Kronecker sets, with \(\varepsilon < \sqrt 2\), if \(\Gamma\) has only a finite number of elements of order two. The Ramsey-Wells-Bourgain \(B_d (E) = B (E)\) characterization of \(I_0\) sets is also proved. Chapter 10 deals with generalizations of \(I_0\) sets and presents a different way to approach the two main open problems. The book ends with three appendices and a list of 18 open problems. It may be pointed out that the reviewer and \textit{H. Queffélec} and \textit{L. Rodríguez-Piazza} [J. Anal. Math. 86, 105--138 (2002; Zbl 1018.43004)] constructed subsets of \({\mathbb Z}\) which are very close to Sidon sets: they are \(p\)-Sidon for every \(p > 1\), \(\Lambda (q)\) sets for every \(q < \infty\), and are sets of uniform convergence, but they are not Sidon sets (actually not Rosenthal sets), and they are dense in the Bohr group.