(1) Adlet Interaction: Refine the adlet fusion scenario to include all types of adlet composition operators and work out their semantics, i.e., how the adlets should be merged/combined/absorbed/split, how they travel back to original site, stay at current site, continue on to other sites, etc.
(2) Virtual Graph Update Algorithm: Work out the virtual graph update algorithm first in example and then in pseudo code. (As a bonus problem or project, finally in real code.)
(3) Retrieval: Work out the retrieval algorithm first in example and then in pseudo code.
(4) Advertising Strategies: The above scenario describes only the aggressive advertising strategy. The reactive advertising strategy where adlets stay put, and the progressive advertising strategy where adlets initialy stay put and then travel, can be investigated so that the general approaches to active document advertising can be worked out.
(5) Visual Interface for Adlets: Visual language for adlet specification and the visualization of adlet interaction can be studied. You can do mockup (using web pages as a MUST) or trial implementation.
(6) Active Index Testbed: In the above scenario, the implementation using active index is also explained, but a lot of fine points remain undefined. You can go deeper to describe in detail how this can be done, i.e., using the MICE to support adlet system implementation. Bonus for trial implementation using IC_Manager. Notice this requires a distributed IC_Manager, which is yet to be implemented. So the trial implementation is single-node only.