wiki:WikiStart

Version 12 (modified by sauermann, 15 years ago) (diff)

--

Welcome to KaukoluWiki!

KaukoluWiki is a Java-JSP-based Semantic Wiki that manages its data by using Semantic Web tools.

The main reason for developing such a Semantic Wiki was the fact that most Web pages lack machine-readable semantics, i.e. means to include the meaning of a certain piece of data in a formalized representation. This is the reason why automated integration of knowledge and reasoning over this knowledge is not possible yet.

A means was needed to create web pages that semi-automatically include semantic information. This was the main reason for creating KaukoluWiki. It is a Wiki that uses methods of the Semantic Web to bridge the gap between Semantics and the WWW.

Main Features

  • RDF backend for page storage
  • Semantic Annotations that link a word of natural language to a resource (an URI) can be created. The intention of this is to create keywords whose occurrence in the Structured Text of a wiki page prompts KaukoluWiki to list up the !URIs they are linked to. These resources then can be used to create RDF triples representing the text's content in a formalized way.
  • Support for synonym pages and including existing texts into another Wiki page, cutting redundancy.
  • Basic RDF(S) viewer, which is yet somewhat unconnected to the rest of the Wiki but would be integrated further in later versions.

Technical Details

Technically, KaukoluWiki is an upgrade of the existing JSPWiki by Janne Jalkanen (http://www.jspwiki.org/). Whereas JSPWiki uses text files as its backend, KaukoluWiki keeps its data in an RDF store.

The Triples are implemented by the Java framework Jena-2.2, and the Quads by the Java framework NG4J-0.4.

Status

The present version of Kaukolu is an ad-hoc research prototype and not ready for daily use. A rewrite is scheduled as part of the Semantic Desktop-Praktikum 2005.

Improvement Proposals

First of all, the Provider (RDFProvider3) finds the ntriples.txt file (i.e. the file which contains all the Web page-related RDF statements, such as text) via a HARDLINK. Obviously, this is bad, but I have found so far no solution to circumvent this. WikiEngine.getBaseLink() does not seem to work for some reason I cannot comprehend.

Secondly, the data handling of Kaukolu is somewhat clumsily implemented. Actually, all the stuff which is meant to be rendered should be handled in Provider.getPageText(…) or Provider.putPageText(…) , but I put much of the work into the JSP files. The intention was to JSP-page-specifically extract RDF information (such as what to render exactly -the generic and the RDFS viewer renders different texts than regular pages - what class the page is, what properties it has etc.). Due to lack of time, I was not able to create a more elegant, Provider-only solution.

Further improvements would be to implement a very basic RDFS “editor”, to introduce onMouseOver or JS elements or to work with some other data structure than just Triples.

separate Kaukolu source from jspwiki? LeoS: The two source branches are at the moment twisted together. It may be wise to move the kaukolu-source to a branch in kaukolu/src-kaukolu/src and have your own paths then: something like:

  • kaukolu/src-kaukolu/src - source
  • kaukolu/src-kaukolu/examples - examples
  • kaukolu/src-kaukolu/test - junit tests

This would allow to keep updated with jspwiki.

Troubleshooting

It is essential that the path to the repository files (ntriples.txt, trix.txt, rdfs.txt) is defined correctly in the Provider and in every JSP file (if applicable) is set correctly. Otherwise, there will be FileNotFoundExceptions thrown.

The path to the default page is <Wiki path>/Wiki.jsp?page=Main. Without a "page" parameter, there could be errors.!

Webstats4U - Free web site statistics
Personal homepage website counter
Free counter