3 - The Origins of Asserted Versioning: Computer Science Research

Abstract

This chapter discusses computer science contributions to temporal data management, and the relevance of some of these concepts to Asserted Versioning. It begins with an overview of the three sources of Asserted Versioning: computer science work on temporal data; best practices in the IT profession related to versioning; and original work by the authors themselves. Over the last three decades, the computer science community has done extensive work on temporal data, and especially on bitemporal data. During that same period of time, the IT community has developed various forms of versioning, all of which are methods of managing one of the two kinds of unitemporal data. Asserted Versioning may be thought of as a method of managing both uni and bitemporal data which, unlike the standard model of temporal data management, recognizes that rows in bitemporal tables represent versions of things and that, consequently, these rows do not stand alone as semantic objects. Asserted Versioning may also be thought of as a form of versioning, a technique for managing historical data that has evolved in the IT industry over the last quarter-century. But unlike existing best practice variations on that theme, Asserted Versioning supports the full semantics of versions. In addition, Asserted Versioning also integrates the management of versions with the management of assertions and with the management of bitemporal data. Besides embracing contributions from computer science research and from business IT best practices, Asserted Versioning introduces three new concepts to the field of temporal data management: episode, the internalization of pipeline datasets, and encapsulation.