SWOOKI Approach
SWooki is the rst attempt to build a peer to peer semantic wiki. SWooki is based
on Wooki [
1
] a peer-to-peer wiki system. SWooki integrates the semantic web
technology by following the philosophy of Semantic Media Wiki[
2
]. The semantic
annotations are embedded in the wiki text via a wiki markup e.g. typed links.
It follows the use of wikis for ontologies approach. A formal ontology emerges
during the edition of the wiki pages. SWooki provides the same functionalities of
any server-based semantic wiki. In addition, SWooki allows the following three
interesting use cases for P2P wikis: (1) a massive collaboration, (2) the o
line editing, and (3) an ad-hoc collaboration [
3
]. In order to combine P2P wiki
system with semantic wiki systems, it is very important to know how semantic
wikis represent their semantic data and how they combine textual parts with
semantic parts. The main issue that we address is how to merge wiki pages that
contain semantic annotations and if this combination changes the behavior of
the Semantic Media Wiki. In this paper, we investigate how we can combine the
Wooki with the Semantic Media wiki. We called this combination SWooki.
We adopt Wooki [
1
] because it supports all use cases for a P2P wiki system
previously cited. In Wooki, wiki pages are replicated over all members of the
p2p overlay network. A wiki page is considered as a sequence of lines. Each
server hosts a copy of pages and can autonomously o er the wiki service. Page
copies at each site are maintained by an optimistic replication mechanism called
Woot [
4
] that disseminates changes and ensures consistency. Woot ensures the
CSCW principles of convergence and user intentions. The only two available
merging algorithms for peer-to-peer wikis are the Thomas rule [
5
] and Woot
strategy. These algorithms handle linear text. The merge using these algorithms
is done by the server which di ers from the merge done by the users during
concurrent editing in centralized semantic wikis. In case of Thomas rule, only
the modi cations of the user that lastly saved are kept. In case of Woot, the result
includes the modi cations of all users. However, the result in Woot is produced by
the server and must be reviewed by a human in order to verify its accuracy. The
Woot algorithm preserves users intentions, all concurrent e ects are visible in the
nal version of the wiki page. If the user wants to change the result of the merge,
he can do that easily. In conclusion, Swooki adopts the Woot algorithm because
it provides the best solution for the merge by keeping all concurrent changes
made by the users without any lost of updates. The SWooki approach allows to
build a P2P semantic wiki very easily by integrating semantic annotations into
a P2P wiki. It allows also to balance the load of queries. It provides a cheap
way to have many replicas of the same semantic wiki. This total replication of
semantic data can be used to distribute semantic queries on di erent replicas.