Referential Integrity of Links in Open Hypermedia Systems

This paper is concerned with broken hypertext links. These are links which do not refer the reader to the information that was intended by the author of the link. The paper presents three distinct models which have been adopted by various developers for the storage of hypertext links, and considers the problems that may result from adopting each of these models, and reviews and classifies a number of methods that may be adopted for preventing these problems. The link models that are reviewed range from the tightly coupled links implemented by html in the World Wide Web, through to the loosely coupled links adopted by some link server systems. The paper concludes that there can be no universal solution to this problem; rather there is a range of approaches from which hypertext developers must choose a solution appropriate to their needs