You're right, I should update the main page for users to get an update on the current progress. For people a little bit more involved though, they can see that the forums are very active and betas are constantly being released. Currently we're at 1.5.9.2, if you set in the preferences, 'update on beta' you can also get the latest.

True multi-source downloading has already been implemented and I'm fine tuning it currently for the final release of 1.6. Hashing provides assistance for me in assuring that chunks are part of the same file even before I start downloading it, but it isnt essential. Which means its on the list, but not a critical feature.

Ultrapeer does two good things. Lowers the network bandwidth requirement for users, and raises the visible nodes for all users on the network. Once the specs are fully hammered out and the limewire client proves it to be nice and efficient, it will be integrated into gnucleus.

Auto-researching in gnucleus works like this. If you are pending a download, have a search running, or downloading a file, then re-searching is done by gnucleus to get you more results. This doesnt mean a query broadcast is sent to all connected nodes on some time interval. How gnucleus does it, is on every new connection it makes, a query is sent only over that connection. This is much, much more efficient, and the results you get are more likely to be different than something you already have.

Thanks for the questions.
John Marshall