Sets with large Borsuk number (Q1580073)

Let \(E\) be a bounded subset of Euclidean space \(R^d\) and \(\delta = \text{diam} E > 0\). The Borsuk number \(b(E)\) is the smallest positive integer \(k\) with the property that \(E\) can be partitioned into \(k\) sets all with diameter less than \(\delta\). For the \(d\)-dimensional ball \(B^d\), \ \(b(B^d) = d+1\); this motivated \textit{K. Borsuk} [Fund. Math 20, 177-190 (1933; Zbl 0006.42403)] to ask whether \(b(E) = d+1\) for every \(E\subset R^d\). The conjectured answer YES became known as the Borsuk conjecture, and its validity for \(d=2, 3\) was established. Nevertheless some mathematicians doubted the validity of the conjecture. Almost 60 years later \textit{J. Kahn} and \textit{G. Kalai} [Bull. Am. Math. Soc. 29, No. 1, 60-62 (1993; Zbl 0786.52002)] proved that the conjecture is false. More precisely they proved: if \(m\) is a prime power and \(d(m)={4m\choose 2}-1\) then there exists a set \(E \subset R^{d(m)}\) with \(b(E) \geq q(m):= {4m\choose 2m}\Big/{4m\choose m}\). (\(q(m)>d(m)+1\) when \(m\) is large.) This shows that the Borsuk conjecture is false for dimension 2015. Other investigators [\textit{A. Nilli}, Contemp. Math. 178, 209-210 (1994; Zbl 0816.52001); \textit{J. Grey} and \textit{B. Weißbach}, Preprint 25, 1997, Univ. Magdeburg; \textit{A. M. Raigorodskii}, Russ. Math. Surveys 52, 1324-1325 (1997)] created counterexamples establishing bounds which lowered the dimension, i.e., showed that the Borsuk conjecture is false for dimension 946, 903, 561 and (in the paper under review) 560. The author comments on this new record with the words of Horace: ``parturient montes nascetur ridiclus mus''. True, but this is a fine paper, concise and clear, interesting and ingenious.