The biggest problem faced by developers of small-scale online commities is to ensure user participation. We apply techniques based on social psychology and visualization to motivate users to participate and we build incentive mechanisms to regulate the amount and quality of contributions.

We have developed a peer-to-peer file and bookmark sharing system, called Comtella, allowing students and researchers to share papers (either their own, or downloaded from the web). Comtella does not aim to replace web-based search engines like Google. Its purpose is to manage bookmarks or documents that people find using those engines and save for future use. Comtella has been used in three undergraduate computer science classes, where students can put together their efforts to find class-related resources on the web. We see also big potential value for research groups, which can benefit from putting together their literature research efforts and create a digital library with zero maintenance costs.

Since 2005, Comtella has evolved into two projects:

  • Comtella-D - a bookmark sharing and discussion forum forum system used as a platform to explore incentive mechanisms based on immediate feedback on ratings, colour/heat-based recommendations, and reciprocal relatioships visualization. It was used in a class on Ethics and IT in 2006.
  • Multi-community Comtella - an infrastructure which allows any user to create and own many communities. This system was used as a platform to study issues of decentralized user modeling. This version of the system is still in use but for a particular purpose and with limited access. Please, let Julita know if you want to try it out!



Project Description

Presentations of the project (PDF):

You could try it out (no longer...sorry) :


User participation is crucial in online communities and peer-to-peer systems where users share opinions, files, bookmarks, services, like compute cycles, or human help. Our experience with I-Help, a P2P system for peer help showed that the existence of a critical mass of users at any time is a necessary condition for the system to be functional. Even file sharing applications, like KaZaA, Napster or Limewire (where files can be reproduced endlessly at no cost), try to ensure user participation by "tricks", e.g. making it difficult to quit the client, by certain default settings for file-sharing which are difficult to change. We believe that when it comes to sharing active services, it will be much more challenging to ensure incentives for user participation. One possible way to ensure such incentives is through introducing an economy and micro-payments. Users do not necessarily need to know about the currency transactions, which will be handled by their agents .

The currency will help regulate the suppy and demand of resources and services. The reward for the user will be in terms of better quality of service. Those who contribute more will enjoy better quality of service than those who don't contribute. Of course, once there is a reward, some people will try to game the system. Therefore the design of motivational mechanims has to be "evil-proof".

Another way to address the issue of motivating participation is through teaching users to be more conscientious and better "netizens". We are working on developing motivational interface that uses positive and negative feedback through graphical images to certain user actions. Modelling of users' interest and social relationships is applied to help visualize the community and the user's social status.


** adaptive rewards mechanism **