Furthermore, Apostolico and Drovandi suggest to use $\pi$-code, see Section 4 from [their paper](https://papers-gamma.link/paper/178), when the power law distribution have an exponent close to 1. It actually pushes me to ask, maybe a naive question: is there any standard method to construct such a code when the distribution of the gaps is given or estimated from the data?