NEPOMUK Deliverable D11.2: Mandriva Community Case Study - First Prototype of A Social Semantic Help Desk

Last modified by slauriere on 2007/10/17 15:43

NEPOMUK Deliverable D11.2: Mandriva Community Case Study - First Prototype of A Social Semantic Help Desk

 First Prototype of A Social Semantic Help Desk
Publication date (last release)15 October 2007
Dissemination levelPublic
Download linkD11.2_v10_NEPOMUK_1st Prototype of Social Semantic Helpdesk.pdf

Executive summary

The objective of WP11000 is the adoption, application and validation of the services provided by the NEPOMUK platform in the context of the on-line community of Mandriva Linux users. WP11000 is designing a prototype providing the members of this community with a new generation tool for sharing knowledge related to the Mandriva Linux project.

Large open-source projects use and produce huge amounts of semi-structured data exchanged through numerous channels such as wiki pages, mailing-lists, forums, issue tracking reports, requirement specifications, task descriptions, commented source code, instant messages etc. This massive amount of information is constantly activated as a resource for knowledge which helps the project partakers to solve existing problems collectively, and which sustains the collective innovation process at stake in the design of the next versions of the projects. This collective process has led to impressive results with relatively simple knowledge management tools, i.e. wiki engines, mailing-list servers, source control patch/diff systems and instant messengers. The Linux eco-system is one of the result of this process, consisting in 2007 of millions of users worldwide, thousands of contributors and of companies doing business around the related platform. A typical Linux user or developer has to deal with a huge amount of information in order to get the most of his system and of the project he's interested in.

By providing customizable tools for linking the mental model of the user with the collective knowledge of the community, the NEPOMUK platform applied to the context of a Linux user's activities makes the process of learning, getting help and contributing to existing knowledge easier and more efficient.

This report describes the first version of a community help desk system combined with a knowledge base engine that is being experimented in particular with the Mandriva Linux community through the Mandriva Club platform. Even though the prototype focuses on help desk aspects, the approach, if successful, could be widened to all the processes involved in the collective design, use and maintenance of a software system by allowing contributors to annotate not only the questions and answers about the system, but all artefacts of the software forge: source code, bugs, specifications, tests etc. In that sense, the NEPOMUK platform is likely to be of interest in the emerging domain of Semantic Web Software Engineering.

As of October 2007, a first version of the WP11000 prototype has been designed and brought on-line as an additional service in alpha stage of the Mandriva Club. This prototype features a tight integration with the NEPOMUK-Eclipse desktop prototype. The system builds on the one hand on NEPOMUK components and on the other hand on XWiki components. XWiki is an open-source extended wiki engine and a generic platform for collaborative Web applications. The prototype's code name is currently XWITS, standing for "XWiki Issue Tracking System": help desks and issue trackers share many common concerns indeed, and the designed system may be experimented later on as a semantic issue tracker.

The first version of the WP11000 community help desk lets the users publish questions and answers and add detailed metadata to them so that the information search and discovery processes get significantly improved compared to traditional help desk systems. Metadata can be added both collectively on the server side or individually on local desktops for the purpose of learning and memorizing. Users can for instance link a question and the set of related answers with some personal files, emails and bookmarks, all together related to a technical task. The established semantic graph is then meant to let him find again information efficiently when faced to a similar task or issue.

The current help desk prototype shows a solid basis which will be extended in the next months according to a development roadmap outlined in this report. Now that the assembling of NEPOMUK components with the XWiki ones has been carried out, it is expected that the next design and implementation steps will occur at a fast pace, with approximately one release per month until the end of the project. The next steps will consist of adding a P2P and a social layer to the system in addition to the semantic one through the integration of the NEPOMUK distributed index and of the NEPOMUK community manager. The semantic layer itself will be enhanced by the integration of the NEPOMUK text analysis service as well as the capability of dealing with the user's context. The help desk will be improved both on the backend and on the desktop. Beside the use of NEPOMUK-Eclipse, NEPOMUK-KDE and NEPOMUK-Mozilla libraries will be put into practice for drawing further the links between a NEPOMUK desktop and the help desk backend, with the objective to create a large scale Semantic Web application tightly integrated with the users' desktops.

Nepomuk - NEPOMUK - The Social Semantic Desktop - FP6-027705

Nepomuk Consortium 2006-2008