UCIAD

1. Data Capture

1.1. Experience in making activity types / patterns and identification patterns emerge from heterogenous logs

1.2. Accurately recognising users across logs - is IP plus user agent sufficient?

2. Algorithms

2.1. Pluggable architecture to handle different log formats into a ? scheme / ontology

2.2. Can ontology alone ? recognise ? (unlikely) and what more is needed?

2.3. How to compile traces of activities that the same but are captured from different systems

3. Scalability

3.1. Experience in loading huge amounts of activity data onto a triple store and reasoning with it

3.2. Scalability of storing, analysis and implementing large amounts of traces of activity (in our case, as RDF)

4. Data Release / Open Data

4.1. Ontologies ? and for recognising ? ? ?