This is something like a textual search engine - only queries are static in the form of article titles, and the answers are pre-computed. Ideally Wikipedia should have a system which would allow answering questions of this form - and trace results to original data. In other words, Wikipedia may have "kernel" articles and "derivative" articles, and be able to generate second from first, both as a compression and as a convenience.
Interesting, but I'm more curious about the robot with the pen.