Automated document revision trimming in a collaborative multi-user document store

摘要

Methods, systems and techniques for managing revisions of documents in a collaborative, multiuser document store are provided. Example embodiments provide an Automated Document Revision Management Server (“ADRMS”) to automatically cluster and remove revisions of file content for easy navigation and management. Revisions are trimmed when necessary to conserve storage space. The ADRMS creates logical clusters of revisions based upon some measure of their similarities. That is, revisions that are similar and can be represented by the latest revision in the cluster formulate one cluster, and those that are markedly dissimilar are placed in a different cluster. The logic used to cluster revisions accounts for time-based factors, content-based factors, and context-based factors to determine whether a revision is incremental and can be grouped in the same cluster or is significant enough to warrant a new cluster. Revisions may be trimmed based upon age and/or available space by a revision trimming component.

主权项

1. A computer-implemented method in a server computing system for trimming revisions of a file to facilitate management of revision growth, the revisions stored in a document store accessed by a plurality of users and received from a plurality of distinct client systems, the revisions previously grouped into a plurality of logical clusters, comprising:
automatically, under control of the computing system,
determining which revisions of the file are an essential revisions of the plurality of logical clusters and therefore represent their respective logical clusters and which revisions of the file are non-essential revisions of the plurality of logical clusters;determining an amount of available space in the document store;designating as hidden a portion of the non-essential revisions randomly chosen across the plurality of logical clusters and deleting a portion of the non-essential revisions randomly chosen across the plurality of logical clusters, wherein the portion of the non-essential revisions that are designated as hidden and the portion of the non-essential revisions that are deleted are determined based upon the determined amount of available space in the document store and age of the determined non-essential revisions; andpresenting the revisions of the file to a user by presenting indicators to each essential revision of each of the plurality of logical clusters and presenting by default only indicators to the non-essential revisions of each of the plurality of logical clusters that have not been designated as hidden or have not been deleted.