On the zero defect conjecture (Q518184)

A palindrome is a word identical with its mirror image. The palindromic defect of a word \(w\) is the non-negative number \(D(w)=| w| +1-p_{w}\) where \(p_{w}\) is the number of palindromic factors of \(w\). The palindromic defect of an infinite word \(u\) is \(D(u)=\sup\{ D(w):w\text{ is a factor of }u\}\). The infinite words investigated in the paper are fixed points of primitive morphisms. A morphism \(\varphi\) is primitive if, for some \(k\geq1\), for any pair of symbols \(a\), \(b\), \(\varphi^{k}(a)\) contains \(b\). Brlek's Zero-Defect Conjecture states that any fixed point of a primitive morphism with finite palindromic defect is either periodic or its palindromic defect is zero. This conjecture was proved to be false for alphabets of size at least \(3\). The paper under review gives the proof that the conjecture is true for binary alphabets, and provides a subclass of primitive morphisms over bigger alphabets where the conjecture is still valid. The proofs are based on the investigation of properties of special bipartite extension graphs associated with factors of an infinite word.