Relationship-Based Dynamic Versioning of Evolving Legal Documents

Abstract

Rule evolution is usually performed by creating a new document which explicitly details changes to specific parts inside other rules’s content. Obtaining (virtual) document versions corresponding to a rules’s state at a specific date is thus left to document users, who manually extract from library collections, and compose, the pieces of text needed to obtain the desired version. When changes are numerous this can be a tedious task. We propose a solution to dynamically generate virtual rule versions on user demand, respecting the library documents integrity. References to other documents and modification relationships can be automatically detected and are modelled as typed links - modelled with XLink - in a relationships graph. This graph can be used to query relationships, to create hypertext, and to dynamically generate rule versions. In this paper, we focus on the version generation process: a dynamic document composition based in a graph traversal, during which we intelligently infer the composition rules of the desired version.