Collaborative editing of JSON objects based on operational transformation

Abstract

Collaboration over the Internet is becoming more important these days. Especially due to mobile devices and online university courses more people want to and have to edit data shared between them concurrently. Operational Transformation (OT) is a group of algorithms, which received a lot of attention in the area of collaboration. The direct application of local changes hides latencies from the users. Still, after exchanging the operations, a consistent result is eventually reached. Most of the applications based on OT were limited to multiple users editing a text document because of the missing theoretical foundations for other data structures. A recent proof published by Jungnickel and Herb extends the theory of OT to n-ary trees. A following publication expanded the thought to JSON objects by suggesting that they can be represented as trees and the tree transformations could be applied to them. This thesis continues this chain of thought by first summarizing the current theoretical base of OT and then formalizing OT for JSON objects. Based on these findings a library was developed that enables OT for three data structures, namely linear ones, trees and JSON objects. The development of the library was done in an iterative manner and described in detail. Supporting the collaborative editing of JSON objects is a distinguishing trait of the library because all software that supported this so far lacked the theoretical foundation. Lastly, the library’s performance was compared to the one of some competitors. The library developed with this thesis is the first one that has a theoretical foundation for collaborative JSON editing. The tests showed that all requirements were fulfilled, a consistent state was eventually reached and that it has a competitive performance. Therefore, the library can enable collaborative features for software, which stores its data as JSON objects.