Open Semantic Hyperwikis

Abstract

Wikis are lightweight, community-editable, web-based hypertext systems, which can be described as a website that anybody can edit. From this collaborative base has grown significant efforts at large-scale knowledge management such as Wikipedia. Recently, ‘semantic’ wiki systems have been developed with typed links, such that the structure of nodes and links is analogous to an RDF graph of resources and arcs: a machineprocessable representation of the relations between articles which can form part of the web of linked data. Despite this, the hypermedia side of wiki systems has so far largely been constrained to the web model of simple embedded, unidirectional links. This research considers the hypertext origins of wiki systems, asks, and answers how the technologies developed during decades of hypertext research may be applied to better manage their document, and thus knowledge, structure. We present experimental evidence supporting the hypothesis that additional hypermedia features would be useful to wiki editors on both macro- and micro-scales. Quantitative analysis of editing logs from a large-scale wiki shows that hyperstructure changes form a substantial proportion of editing effort. Conversely, qualitative user studies show that individual user editing can be better supported by classical but since overlooked hypertext features such as first-class links and transclusion. We then specify an extensive model for a ‘open semantic hyperwiki’ system which draws from these fields, based around first-class links with support for transclusion and advanced functional link types, with defined semantics for the role of versioning and parametric nodes in the linked data world, while mindful to preserve the core simplicity that allows non-expert users to contribute. This is followed by a practical approach to its implementation in terms of an existing experimental modular wiki foundation, and the actual prototype implementation, which has been made available as open source software. Finally, we work through applying the system to a set of real-world use cases which are currently employing classic, non-semantic wiki software, and evaluate the implementation in comparison to a conventional semantic wiki in a user study.

More information

Identifiers

Catalogue record

Export record