598trustbased P2P exchange on the semantic web and the WYMIWYG KnoBot project

An agent for decentralised knowledge exchange

Reto Bachmann-Gmuer
Language german
Room Saal 1
Time Day 3, 13:00h
Duration 2 hours
[1] Preliminary System online



An agent for decentralised knowledge exchange

Table of Contents

<#summary>1. Summary

<#abstract>2. Overview

<#description>3. Description of project

<#id352247>4. Description of technology

<#id352495>5. Description of user groups and deployment

<#id352503>User categories

<#id352512>Early adopters

<#id352529>Focus group

<#id352557>Information poor

<#id352579>6. Description of civil society application for project

<#id352615>7. Description of team

<#id352754>8. Timeline/Suistainability


<#id353211>9. Further resources


KnoBot combines semantic web technologies and a radically decentralised approach with the insight that the challenge of infomation selection is at least as important as infomation delivery. KnoBot addresses the issue of how to get the most relevant informmation by passing data along the paths of established trust relations. The way infomation is forwarded and wheihted in informal social networks gets its electronic counterpart.

WYMIWYG is working on java based implementation of KnoBot. WYMIWYG is a non commercial association open to new participants.


Information overflooding, monopolisation, centralisation and surveillance are the current problems of the proclaimed knowledge and information society. The crucial challenge is getting the relevant information out of a chaotic ocean of data, KnoBot deals with this problem by combining semantics and trust-based p2p systems and frameworks.

KnoBot allows to position articles in a space of interrelated topics. It overcomes sharp categorisations[1] to allow ordering information in a fashion that is closer to human memory. While it can be used standalone as a kind of mind mapping tool or expose its content over the web like a blog, its full potential arises when it can communicate with peers. A network of KnoBots form a socio-topical space that provides information that is highly relevant and trusted for the user. This is done by establishing relationships of trust between users and of similarity between topics and spidering the resulting network.

[1] People tend to have a strong wish to categorize things sharply. This makes sense for certain applications, such as a book that should be on a specific place in a library. But development of language and knowledge depends on ambiguity and unsharpness, fuzziness.

Chapter?3.?Description of project

Garbage comes first, then we build a system to deal with it (Don DeLillo)

The WWW is a huge, unstructured and hardly mapped pile of content. The individual wanting to gain knowledge from the WWW is challenged by the impossibility of universal or constant relevance criteria. The dependence on search engines with nontransparent algorithms, the inequality of power and the factual importance of informal information selection authorities foster anonymous consumption rather than active knowledge-creating exchange. Most people access news through information portals built by well known media brands. Having a "central" position on the WWW correlates much more with the financial power than with the quality of the information offered. This slows down the growth of knowledge and makes censorship easy.

The problem is confusion, not ignorance (Karl Weick)

KnoBot is a knowledge sharing software that builds a network of topics that expands over people. It bases on trust-relationships between persons and relationship of similarity as well as of sub- and superordination of topics.

A person using KnoBot may enter articles herself and (optionally) associate them to topics she defined. The person may read RSS[2]-feeds and associate the feeds or individual articles to her topics.

Articles can be associated (article-topic relation) to one or several topic with a specified relevance (r). Furthermore, topics can be linked (topic-topic relation) to other topics, with a relevance factor (f). An article associated to a topic T1 also has inferred associations (with relevance r*f) to all topics Ti that are linked to T1 and so on.

KnoBot exposes the topics as enhanced RSS feeds with or without password and encryption. These feeds are accessible with standard RSS aggregators but accessing them with KnoBot offers the following advantages:


Client can specify down to what level of relevance they want to retrieve articles.

* *

The user can make topic-relations to topics of other users and take advantage of the their relevance ratings and topic network.

* *

Clients don't retrieve the same article twice but get only new article (with a relevance higher than the specified minimum).

* *

Clients can access encrypted and password protected feeds.


The first persons using KnoBot will be those that write blogs already now. KnoBot offers the possibilities of common blogging software such as "Movable Type"[3] and makes it easy to syndicate and republish remote RSS feeds. Using KnoBot bloggers can do what they already do, but with the KnoBot - Network they will continuously discover articles their friends or friends-of-friends found interesting in their blogs.

KnoBot will be released under Apache-Style license and if possible dual licensed with GGPL[4] (this will be checked when the final version of the GGPL is released).

[2] http://www.mnot.net/rss/tutorial/ explains RSS

[3] See http://www.movabletype.org/

[4] See http://www.ggpl.org/

Chapter?4.?Description of technology

KnoBot relies on the following technologies:


Jena Semantic web framework

Jena is a Java framework for building Semantic Web applications. It provides a programmatic environment for RDF, RDFS and OWL, including a rule-based inference engine. See http://jena.sourceforge.net/

* *

Java Servlet/Filters technology

A main component of KnoBot is the RDF Web Component Framework (WYMIWYG-RWCF). RWCF extends the java servlet and servlet-filter technology to allow access to a common RDF repository. KnoBot heavily relais on filter for modularization.

* *


The HTTPSY URI scheme[5] extends HTTPS to allow secure communication without the requierement of trusted certificate authorities and without relying on the DNS System for authentication.

* *


The Swiss Refugge Council has release tools for editing and rendering RSS 1.0 feeds which base themself on WYMIWYG-RWCF

* *

RSS 1.0[7]

The exchange of content bases on RSS 1.0 and existing as well as new modules for it, the new modules are documented and suggested on the RSS-Dev-list[8]

* mod_content[9] * *


WYMIWYG has already proposed modification for this module to guarantee compatibility with RDF-based aggregators.

* * mod_syndication[11] * *


A module to be introduced for expressing the relevance of an item for a channel


* *


The friend-of-a-friend network and ontology will be used for discovering new sources of information. Extension for expressing trust relationships will be proposed.


[5] See http://www.waterken.com/dev/YURL/httpsy/

[6] See http://development.sfh-osar.ch

[7] See http://web.resource.org/rss/1.0/

[8] rss-dev@yahoogroups.com

[9] See http://web.resource.org/rss/1.0/modules/content/

[10] See http://web.resource.org/rss/1.0/modules/link/

[11] See http://web.resource.org/rss/1.0/modules/syndication/

[12] See http://www.foaf-project.org/

Chapter?5.?Description of user groups and deployment

Table of Contents

<#id352503>User categories

<#id352512>Early adopters

<#id352529>Focus group

<#id352557>Information poor

User categories

Early adopters

Bloggers and independent media centers will be the first to adopt KnoBot. For them KnoBot is just a user friendly and highly standard compliant and free software for doing what they already are doing.

Focus group

Human right activists, media activists and geeks constantly deal with a vast amount of information. KnoBot will help them get the most relevant information for their knowledge context and defined purposes.

Thanks to its strictly decentralised approach it is resistant against censorship. The trust based information exchange makes it safe against information overflooding attacks. Therefore it will be of particular use in authoritarian and repressive environments.

Information poor

While KnoBot was designed to help extracting knowledge from an ocean of information, it can also be used to get the most out of limited bandwidth by reducing the data to relevant information. In fact, it makes little difference whether the interface between computer and human or a network link is to be relieved.

Chapter?6.?Description of civil society application for project

KnoBot makes it easier to get in touch with people dealing with similar issues and sharing similar solution strategies. This will increase the chances of the emergence of new critical public spheres and social movements. These movements will be characterised by a low level of formalisation and little energy investment in institutional reproduction compared with established models of organisation in favour of faster spontaneous objective driven (re)constitutions.

The hope is that KnoBot will widen the social and political participation in a basically egalitarian and intrinsically non-violent fashion by loosening the binding of reproduction of social realities with existing hierarchies and patterns.

Chapter?7.?Description of team

Although we acknowledge the importance of team-work, and several people already have been involved at this rather early stage of development and testing, and it is beyond doubt that even more people will contribute their talent as the project moves forward, grows and gains traction within the free/open source software developer community, the person charged with developing the system, as far as its current form is concerned, is Reto Bachmann-Gmuer. Reto is a Swiss semantic-web hacker with a background in the development of Web-Application since 1996. He worked on the open source content management system mir[13] and was a speaker in the semantic web track of OSCOM 3[14] conference. He contributes to RDF based technologies such as FOAF, RSS 1.0 and Annotea. He is currently part-time employed at the Swiss Refugee Council[15] where he works at OSAR CMS.

Nevertheless, a system's usability requires more than the programming ingenuity that Reto has handsomely offered, and to this end, other people, including George Dafermos, have coalesced around the project. More specifically, George has been testing the system at this phase, and also have been helping with design and information architecture issues such as which features should be implemented in future versions, which features function as anticipated, and in overall, providing the main developer with an experienced end-user's perspective who knows what he needs from such a system.

George is a long-time free software/open source enthusiast, activist, author, and researcher focusing on the intersections of technology, business and society. A frequent speaker at IT conferences, George has delivered presentations to audiences as diverse as the Oekonux Project in Germany, the Berkman Center for Internet and Society in Cambridge, USA, the Aizu University in Aizu-Wakamatsu, Japan, the Open Source Institute in Tokyo, Japan, the Libre Software Meeting in France, and the Chaos Communication Camp. In addition, George is involved in several Internet-enabled free/open source projects, such as the 'Free People - Free Methods' project hosted by Libroscope in Paris, France[http://www.libroscope.org], the Greater Good Public License [http://ggpl.org], GnuBook [http://gnubook.org], and Organis [http://ggpl.org/organis], all of which are premised on workplace transformation through clever economics. In parallel with pursuing his free/open source centric interests, George consults organisations on all aspects of management of innovation and technology. George is currently located in Heraklion-Crete, Greece.

Manuel Henseler has contributed to the design with his perspective as neuroscientist. He helped understanding the complexity of the KnoBot network. His vision is the integration of self-learning system properties as they are used in the research field of neuroinformatics which could further increase KnoBot's potential. Manuel is located in Zurich, Switzerland

Stephanie Hering has recently joined the team analyzing the possible social impacts of KnoBot. She has a background in sociology and political science and will especially promote KnoBot in an academic and socially engaged field. Stephanie works at the university of Bern, Switzerland.

Responsible fiscal agent for KnoBot is WYMIWYG[16] an international not-for-profit association according to article 60 and the following articles of the Swiss Civil Code. The purpose of the association is to develop open source software with social impact.

[13] http://mir.indymedia.de

[14] http://www.oscom.org/

[15] See http://www.sfh-osar.ch/

[16] http://www.wymiwyg.org


Table of Contents





* Most of the features of typical blogging software are implemented.* * KnoBot has been presented on relevant mailing-lists.* * The software runs out of the box, no installation needed.* * Quick start tutorial is available that explains how to perform basic customisation.* * Community website for developers and users is running.*

* *


* The software has been tested as blogging software.* * The feedback has been categorised and key-bugs and features are defined.* * A technical documentation of the project is available.* * A user tutorial is available.* * The decentralised exchange functionality is available without encryption.* * RSS and FOAF extensions have been officially proposed.*

* *


* The key bugs are fixed.* * The user documentation is almost complete.* * A small network of KnoBots is operational.*

* *


* The experiences with the KnoBot network are evaluated and necessary changes identified.* * The software has been presented at conferences in the field of open source and independent news media.*

* *


* Encrypted communication is implemented.* * The user interface has been simplified.* * User and developers documentation is complete.* * The website of the project has been enhanced for a wider audience.* * Possibilities for donations and paid adaptations are advertised.*

* *


* The software has been presented in relevant medias.* * The number of developers has grown.*

* *


* The digital communication between KnoBot developers is now exclusively based on KnoBot.*

* *


* A activity plan for 2005 has been written taking into consideration the donations that has already been received.*

* *


* Lobbying work has been done to promote KnoBot and as such increase the value of the network.*


Chapter?9.?Further resources

Presentation of KnoBot at OSCOM: http://oscom.org/Conferences/Cambridge/Proposals/bachmann_knowledge_exchange.html and http://slideml.bitflux.ch/files/slidesets/539/title.html

Preliminary System online: http://george.wymiwyg.com

Archived page - Impressum/Datenschutz