Разработка многопользовательской модели данных с использованием подхода Operational Transformation

Abstract

Providing users with the way to collaboratively edit shared model is the challenge of complex informational systems. The majority of existing methods require online connection between users, otherwise, users’ intentions can be lost after a certain delay. The aim of this research is to build a collaborative system that allows users to edit shared model in offline manner without losing their intentions. Firstly, commit-based system was selected as underlying data storage. Secondly, we examined existing collaborative algorithms and selected operational transformation as the most perspective for this scenario. Thirdly, we described the way to merge commits, that stores operations, using modified operational transformation algorithm. Key features of modified operational transformation algorithm include ability to handle resolving of merge conflicts, when intentions of two users are conflicting. To implement this, we consider the set of commits branches as ordered set. As a result, we developed a version control system, that stores user’s intentions and can automatically resolve conflicts on merge. This approach can be applied to any informational system, which is able to save user’s input as the set of operations.