A note on the paper ``On Brlek-Reutenauer conjecture'' (Q442116)

The defect of a finite word \(w\), denoted as \(D\left( w\right) \), is the difference between \(\left| w\right| +1\) (the maximum possible) and the actual number of distinct palindromic factors of \(w\). The defect \(D\left( \mathbf{u}\right) \) of a one-way infinite word \(\mathbf{u}\) is the supremum of the defects of all its factors. Let \(\mathcal{C}_{\mathbf{u}}\left( n\right) \) and \(\mathcal{P}_{\mathbf{u} }\left( n\right) \) denote the number of all factors and the number of palindromic factors of \(\mathbf{u}\) of length \(n\), respectively. If \(\mathbf{u}\) is periodic, \(\mathcal{P}_{\mathbf{u}}\left( n\right) +\mathcal{P}_{\mathbf{u}}\left( n+1\right) =2\). An important characteristics of \(\mathbf{u}\) seems to be the quantity \(T_{\mathbf{u}}\left( n\right) =\mathcal{C}_{\mathbf{u}}\left( n+1\right) -\mathcal{C}_{\mathbf{u}}\left( n\right) +2-\) \(\mathcal{P}_{\mathbf{u}}\left( n+1\right) -\mathcal{P} _{\mathbf{u}}\left( n\right) \). If the set of factors \(\mathcal{L}\left( \mathbf{u}\right) \) of \(\mathbf{u}\) is closed under reversal, then \(T_{\mathbf{u}}\left( n\right) \geq0\) for all \(n\geq0\). Brlek and Reutenauer conjectured that if \(\mathcal{L}\left( \mathbf{u}\right) \) is closed under reversal, then \(2D\left( \mathbf{u}\right) =\Sigma_{n=0}^{\infty }T_{\mathbf{u}}\left( n\right) \). They proved the conjecture in the case when \(\mathbf{u}\) is periodic. The paper [\textit{L.~Balková, E.~Pelantová} and \textit{Š.~Starosta}, ``On {B}rlek-{R}eutenauer conjecture'', Theor. Comput. Sci. 412, No. 41, 5649--5655 (2011; Zbl 1247.68202); corrigendum ibid. 465, 73--74 (2012; Zbl 1254.68193)] extended the validity of the conjecture to uniformly recurrent words, i.e. infinite words \(\mathbf{u}\) where each factor occurs in \(\mathbf{u}\) infinitely many times, but has only finitely many return words. A return word of \(w\) is a word \(q\) such that \(qw\) is a factor of \(\mathbf{u}\), \(w\) is a prefix of \(qw\), and \(w\) occurs in \(qw\) exactly twice. Unfortunately, there was an error in the proof of the main theorem, as shown in the current paper, where a counterexample is constructed. For the correction and the proof of the general version by the original authors, see [Zbl 1254.68193; \textit{L. Balková, E. Pelantová} and \textit{Š. Starosta}, ``Proof of {B}rlek-{R}eutenauer conjecture'', CoRR abs/1208.2505].