The paper studies a specific type of mathematical walk defined as follows: you start with a number, append some digits to it in a chosen base (like base 10), and the result must satisfy a rule given by a linear recurrence relation. A linear recurrence is a rule where each term in a sequence is a fixed combination of previous terms, like the Fibonacci sequence where each number is the sum of the two before it. The question is whether you can keep doing this forever, building an infinite sequence of numbers, each obtained by tacking digits onto the previous one while obeying the recurrence rule.
The key condition involves whether the dominant growth rate of the recurrence, called alpha, is "multiplicatively independent" of the base b. Two numbers are multiplicatively independent if neither is a power of the other in any rational sense. For example, 2 and 3 are multiplicatively independent, but 4 and 2 are not. The paper proves that if alpha and b are multiplicatively independent, then you cannot sustain an infinite walk if you insist on always appending the same fixed number of digits at every step. In other words, strictly uniform digit appending is impossible under this independence condition.
However, if you allow the number of appended digits to grow over time, the situation flips. The same independence condition actually guarantees that an infinite walk does exist. So the paper draws a sharp contrast: rigid, fixed-length digit appending fails, but flexible, growing-length digit appending succeeds. This result connects ideas from number theory, combinatorics on words, and the theory of linear recurrences, and it contributes to understanding how algebraic properties of numbers interact with their digit representations in different bases.