An Architecture for Logging and Searching Chat Messages

Instant Messaging has become an important means of communication between people around the globe, providing an alternative to telephone and email conversations. The number of users using Instant Messenger products has been increasing over recent years. Messenger services provide a perfect environment for private and personal chat. People of all ages log into messenger service or chat rooms to spend time chatting with known and unknown persons. As a form of entertainment for children, there may be danger lurking from adults chatting with teens, camouflaged as teens. This has created a desire to monitor chat activity on a system. The goals of this thesis are to provide an archival system for a chat client, an incremental indexing system and a retrieval system that supports a variety of queries. The archival system is designed to be independent of the messaging software so it can be adopted more widely. The indexing and retrieval systems are portable to both Windows and Linux with minimal changes and the system is to be scalable to continuous operation. Chat sessions can be archived on either the server side (chat rooms) or client side (Instant Messengers). This thesis has focused on client applications, but the development has been coordinated with a related server side project. One of the popular chat clients, Microsoft MSN Messenger, has been chosen to provide the text streams for archiving. The text is indexed and a search capability is incorporated. The retrieval system can support queries based on keyword, username, date, and other criteria.