this is an ambitious project, made of roughly 3 parts:

• structured data templates where all the information should be stored;
• distributed network dispatcher that could send matched information around and
• a browser capable of displaying and inserting all the data.

Storage should be made on a distributed cache fashion, not letting rare information disappear and replicating to more available peers the most popular information. On the user interface it should be able to rate all information and present it to users in a personalized way, where the ratings of trusted people have more weight on a automated suggestion for a certain category of items. Such complex relative matching may become too cpu intensive, so this calculation should be as distributed as possible; automation agents should be implemented in such a way that their ratings should be like phantom ratings of the users they represent, all this in a way that could be validaded or ignored by the user in the most convenient way possible; preferably something implicit from simple movements of her eyes (or mouse, actually).

examples of useful matching could range from:

• insert buy interests and see feeds of sell ads, or vice-versa,
• as well setting up exchanges based on a real or arbitrary monetary system;
• insert personal characteristics and interests and see romantic matches;
• give and receive suggestions of interesting things spread through a subject-oriented the trust network;
• automatic playlist generation, movies suggestions, commented news aggregation;
• input dates and schedules of anything and have it verified and adjusted until fit reality

Lots of time helpers, calendars, date operations, and alike, may help to place items and events on the timescape (as well on the geographic space, with google maps, for example);

• this may help to find people online,
• detect the time peers would be unavailable,
• know the best time to arrive to a party,
• when to go to the bus stop,
• know the frequency any uncommon event occur by collaborating empirical data to form a good estimate of next events.