Whenever something changes in the network (e.g. a new user joins, a user changes their settings on skills and interests, or a user changes location), the matching algorithm is run to compute the set of matching users. The following flowchart shows how the algorithm works.

What follows is a simplified version of the actual algorithm.

 Further Thoughts

In order to keep consistency to infer relations between the user’s skills, a semantic ontology could be defined. This would create a structure to relate activities which are related by a number of possible connections, such as: hyperonymy, meronymy. An example of related activities is: piano – music.

Natural Language Processing tools might also be used to find, categorise and traverse the ontology. The Wordnet lexical database could be employed to lay out the structure frame for the ontology.