Massive collaborative editing becomes a reality through leading projects such as Wikipedia. This massive collaboration is currently supported with a costly central service. In order to avoid such costs, we aim to provide a peer-topeer collaborative editing system. Existing approaches to build distributed collaborative editing systems either do not scale in term of number of users or in term of number of editions. We present the Logoot approach that scales in these both dimensions while ensuring causality, consistency and intention preservation criteria. We evaluate the Logoot approach and compare it to others using a corpus of all the editions applied on a set of the most edited and the biggest pages of Wikipedia.