Abstract :
Seeking answers to questions is a natural part of our learning and social interactions. Although search engines, web-based forums, and inquiries to friends via e-mails or instant messengers are all methods we use today, in many cases, much time is still spent to search, organize, or wait for responses. If knowledgable people can be found online to answer questions in real-time, then the time spent to browse webpages, wait for forum responses, or ask multiple people may be much reduced. In this position paper, we propose Connet, a peer-to-peer (P2P)-based people search system that helps people to have questions answered in real-time by knowledgable contacts via people´s collective social networks. Users share successful experiences of finding responders and achieve greater efficiency when seeking answers to everyday questions. Connet relies on similarity measures between historic and new questions asked, and the trust among mutual friends, to provide more timely question-answering mechanisms. Meanwhile, it also enables a new type of online interactions with the friends of friends.
