Currently, the Retriever site(s) have to check for new content every cron run. These should be frequent so there is little delay in fresh content arriving, but this also means that most of the time the Retrievers are all checking the Distributor site pointlessly.

An alternative method might be this:
- Distributor knows which nodes are of interest to which Retriever
- when a node is updated, Distributor pings the relevant Retrievers
- the Retrievers then run their content retrieval system and get the new content.

This is something I plan to work on in a 6--2 branch. Suggestions and help welcome :D