Semblog: Personal Publishing Platform with RSS and FOAF

Ikki Ohmukai*1, Hideaki Takeda*2

*1 Graduate University for Advanced Studies

*2 National Institute of Informatics

1. Introduction

We propose a personal knowledge publishing platform called Semblog with Semantic Web technologies and Weblog tools. Semblog suite provides integrated environment of information gathering, authoring, publishing and personal networking seamlessly to exchange their own information and knowledge with easy and casual fashion [Ohmukai04].

We designed extension modules for RSS and FOAF, and implemented two types of RDF aggregator called RNA and glucose, which enable to clip and post a Weblog contents. We also developed personalized search and recommendation methods on the Semblog platfor m.

2. Semblog Platform

We propose a content distribution support system for individuals with Semantic Web techniques. We should consider that the information distribution process does not mean only publishing but integrated process containing information gathering and author ing. In our approach, we use a lightweight metadata format that is RSS1.0 [RSS1.0] to activate those information activities. A number of Weblog sites already publish RSS by its content management system, and several applications and ser vices called RSS aggregator are provided based on this trend.

Fig.1 shows system architecture of our Semblog platform based on 4-layers model. First we integrated RSS and FOAF by several extension modules designed for Weblog. These metadata are maintained by Weblog content management system (CMS) like Movable Typ e [MT] and our FOAF TrackBack system described later. We also developed RSS and FOAF aggregator called RNA and glucose which support information distribution process such as collection, creation and publication. Search and recommendation applications are performed on contents/human network constructed by Semblog system.

Fig.1: System architecture of Semblog platform.

2.1 RNA: RSS Aggregation Service

RNA is an extended RSS aggregator described with Perl CGI. Fig.2 shows a snapshot of RNA. The user puts this script to her/his own web server and operates it. Basic function and interface of RNA are shown below.

Fig.2: Snapshot of RNA.

2.2 Glucose: Standalone RSS Aggregator

Glucose is also an extended RSS aggregator but a standalone program for Windows. Fig.3 shows a snapshot of glucose. Different from orthodox aggregator, glucose is developed to support information distribution process by coordinating with RNA. Main func tions and interfaces of glucose are shown below.

We distribute RNA and Glucose in our web site (http://www.semblog.org/wiki/?en). About 2,000 users downloaded RNA and over 20,000 users downloaded glucose from September 2003.

Fig.3: Snapshot of glucose.

2.3 Applications on Semblog Platform

Using functions on Semblog platform, we develop personalized recommendation and retrieval system.

Each RNA has XML-RPC interface that can send and receive its data dynamically. We use Friend Of A Friend: FOAF metadata to identify each RNA. The current version of RNA can generate FOAF data. RNA also has an interface for FOAF management to extend soc ial network easily. We call this method as "FOAF TrackBack".

First the user X enters an RNA URI of the user Y in her/his own FOAF manager. The manager X asks the manager Y to acquire the FOAF data of Y, and writes "X knows Y" link in its FOAF. The manager Y records "Y isKnown X" link in its F OAF and notifies to the user Y. If the user Y agrees, her/his manager registers "Y knows X" link. Repeating this process, a personal network of the user is constructed. We apply several recommendation methods in the network [Hamasaki03] [Ohmukai03].

3. Bottom-up Personal Ontology

In this section we propose a bottom-up personal ontology framework using RSS and FOAF metadata. It is difficult to organize practical ontology hierarchy with top-down approach because building and maintaining such well-organized large ontologies takes a lot of efforts. We aim to develop loose and bottom-up ontology system by combining personal classification, because we consider that personal knowledge will be represented with a routine work such as categorization. Fig.4 indicates a conceptual architec ture of the personal ontology framework.

Fig.4: Personal ontology framework.

Personal ontology metadata consists of FOAF, RDFS Ontology and Contents RSS. The FOAF describes personal information, and the RDFS ontology shows a structure of the categories, and the contents RSS shows written and collected contents by the user.

We add two elements to basic FOAF model. One is <foaf:interest> which is to point the contents RSS, and the other is <rs:personalontology> that is originally defined by our Rough Semantics project (http://www.roughsemantics.org/) to indicate the RDFS ontology. The RDFS ontology is described with the form of Open Directory RDFS format.

The contents RSS is similar to a conventional RSS. Our RSS uses <foaf:topic> to point a category on the RDFS ontology, while the conventional model applies <dc:subject> to express a thesis of a content. This makes our RSS to have backward compatibility. It should be noted that topics pointed by this tag are not restricted to those in their own ontology, rather any topics in others' and some global ontologies. Separating ontology and instances provides flexible manageme nt of each metadata.

Our framework enables applications and services to produce new types of search or recommendation. For example, mapping methods between two directories or bookmarks are applicable to the personal ontologies. Egocentric search is also able to be reali zed easily by building a social network with <foaf:knows> in the users' FOAF.

It is able to calculate a similarity among a personal ontology and the global ontologies such like WordNet and ODP in advance. Multiple personal ontologies can be matched each other via the global ontologies and this method needs less computation co st as shown in Fig.5.

Fig.5: Bottom-up ontology.

4. Conclusion

In this paper we propose a personal publishing system with Semantic Web techniques and Weblog tools. We use a lightweight metadata format like RSS to activate the information flow and its activities. Our system called Semblog platform consists of two t ypes of extended content aggregator and information retrieval/recommendation applications. The system will support not only content publishing process but also information gathering and authoring processes synthetically. Finally we proposed bottom-up pers onal ontology system for richer information distribution.

References

[Ohmukai04] I.Ohmukai, H.Takeda, K.Numa, M.Hamasaki and S.Adachi. Metadata-driven Personal Knowledge Publishing Suite. Proceedings of the Third International Semantic Web Conference (ISWC2004), 2004. (to appear)

[RSS1.0] RDF Site Summary (RSS) 1.0. http://web.resource.org/rss/1.0/spec

[MT] Movable Type. http://www.movabletype.org/

[Hamasaki03] M.Hamasaki and H.Takeda. Find better friends? - re-configuration of personal networks by the neighborhood matchmaker method -. The International Workshop on Semantic Web Foundations and Application Technologies (SWFAT) , 2003.

[Ohmukai03] I.Ohmukai and H.Takeda. Social Scheduler: A Proposal of Collaborative Personal Task Management. Proceedings of Web Intelligence (WI2003), 2003.