Abstract: There is disclosed in an embodiment for management of data for analysis of diverse, multi-structured data from diverse sources importation of data having existing data specifications from an external system. Permissions associated with the data and data policies intrinsic to the data are enforced and any tainted data is untainted by ascribing missing specified properties to the data. The properties for the data that were tainted are recorded and properties of resulting untainted data are encoded as composable metadata. The untainted data and metadata are stored, and recorded tainted properties for the data and/or metadata are tracked. Input of an analysis request is accepted, and the data is analyzed in the presence of the stored metadata. Copies of imported data and metadata are retained and new versions are created when such data or metadata is mutably updated. Results of the analysis are output. Other embodiments are also disclosed.

Abstract: A collaborative editing system described allows multiple collaboration participants to create and edit content in real-time and in a way that allows the content to converge to a desirable intermediate state. In addition, the system supports large content teams in which many collaboration participants use the system at the same time. The system distinguishes between real-time collaborative presentation and traditional real-time collaborative editing. Collaborative presentation occurs when the system displays a first collaborator's changes to a second collaborator, without altering the content that the second collaborator is working on. In this way, the second collaborator is aware of the first collaborator's changes, but the changes of other participants do not directly impact the second collaborator's work. The collaborative editing system also separates participants by task so that not all content modifications are shared with all participants.