The Web in its current state cannot easily be interpreted by machines, as the majority of data resources and APIs are documented solely in natural language. This problem can be solved by augmenting resources on the web with semantic metadata; forming the so-called Semantic Web. An open problem is finding a generic approach to process such data in a streaming fashion and propagating updates to data reusers. An idiomatic solution to this problem should build on top of established Semantic Web standards, such as HTTP and RDF. This paper focuses on getting updates to the end-users as soon as possible, for which the applicability of existing technologies has been investigated. Our findings show that this subject can be further divided into four subtasks: (i) formatting, (ii) versioning, (iii) delivery, and (iv) caching. Although every subtask can be solved using existing technologies, none are sufficient to solve the larger problem. We conclude that the solution must lie in the combination of these existing technologies, and that future research should focus on end-to-end solutions -- and not only the smaller subtasks.
- Andreas De Witte
- Pieter De Clercq
- Robin Devos
- Sieben Veldeman
- Harm Delva
- Dwight Van Lancker
- Pieter Colpaert
Faculty of Engineering and Architecture, Ghent University IDLab, Department of Electronics and Information Systems, Ghent University – imec