Could this be rebuilt for, or ported to, app.net? Seems like a good place for the author to continue what is obviously a project he cares about.
Of course, they own the field, so they could change the rules at any time, thus (once again) killing the project.
From a technical basis, 90% of the effort was building a robust system atop Twitter's. The actual core behavioural categorisation engine was relatively trivial.
Ie, it would require a major rebuild, without lessening the risks at all.