Using Semantic Wiki as a Semantic Web Workbench/Comment-2

From Tetherless World Wiki

Using Semantic Wiki as a Semantic Web Workbench/Comment-2 [ Edit ]
About Page	Using_Semantic_Wiki_as_a_Semantic_Web_Workbench
Title	Poster Description
Commented by	User:Baojie
Date	2008/06/19 11:45:18 AM EDT
Topic:

Content:
See below

Some people argued that semantic web is only a dream. This maybe partly because of its complexity. To use semantic web and to design semantic web applications in the traditional way is not an easy job. One has to learn a long list of technologies, such as RDF, OWL, SPARQL, ontology editors like Protege, ontology reasoners like Pellet, programming APIs like Jena, some programming languages like Java or PHP, and last but not least, description logics. It takes quite a lot of time.

We are trying to reduce the high learning curve of semantic technologies for both end users and application developers, by using semantic wiki.

Probably you are already familiar with normal wikis like Wikipedia. Semantic wiki is an extension of normal wiki by giving each wiki page some annotations. Those annotations can be easily produced by end users, without requiring any knowledge of ontology languages, nor using an ontology editor. A user can simply fill out a form, or for some cases, copy and paste an existing structured data like bibtex into the wiki, and the wiki will do all other jobs to generate semantic annotations. For some users that know wiki script, they may also directly add semantic annotations using a very simple wiki script extension. Using those semantic annotations, semantic wiki can do much more than normal wiki. We just mention two points here. First, it allows automated data propagation [like populating the publication list of a personal homepage]. You can see, on Jim's page, this paper shows up. However, nobody put it here, this information is propagated here because in the paper's page, it says that Jim is one of its authors. Similarly, on Jen's page, this paper also shows up automatically. In fact, every person on this wiki gets a homepage for free. That will reduce quite a lot of trouble in maintaining personal homepages. Second, semantic wiki supports smart data integration [like recognizing all Jim's papers published in difference names] For example, on this page, information about papers, about projects and about person are integrated; we say that it's smart as compared to a normal wiki in that it can query and so some inference. For example, Jim may publish his papers in different names, but since we declared them to be Jim's aliases, all those papers can be inferred to be Jim's papers and show up here. Of course there is much more semantic wiki can do, those are just two simple examples.

Semantic wiki not only helps end users, but also helps application developers. Our work here at RPI is focusing on using semantic wiki to do lightweight concept modeling and to build more tools on the top of wiki. Semantic wiki is used more like a workbench or a programming platform to design other tools than just an application. It is quite powerful and building applications on the top of it is much easier and faster than in the traditional ways.

We have tried to do RDF modeling, as wiki annotations are essentially RDF data; Relational modeling, which is much like building a database on top of the wiki; Object modeling, which allows us to organize information by abstraction and grouping related things together; and Rules like "if something happens then do something, otherwise do something else". Semantic wiki is surprisingly expressive to support such modeling scenarios, given its simplicity in design.

Some of the prototypes we have built are listed here: Semantic Map, like the new online RPI Map; Semantic Blog - it took me 10 hours to build it and now I use it heavily by myself; Semantic Calendar, such as our group calendar; Task and Issue Tracker - it is to track problems and tasks in our group; Literature Repository - it helps us to manage hundreds of papers using a topic ontology; Scientific Writing - we have integrated a latex writing module on the wiki, so we know who will write what paper about which topic, and when it is due; and multiple authors can work on the same paper simultaneously.

There are definitely more power of semantic wiki to be explored and realized. Work in progress includes building semantic email, ontology repository, semantic resume and CV, semantic Call For Papers, semantic conference management, and much more. The best part of it is that, while you might need to learn some wiki script, its quite minimal compared to the whole set of languages and tools you have to learn to build a semantic web application in the traditional way.