> - Tagging "Handicap"
> You'll note that Engadget and Robert Scoble currently dominate the top
> tags ('gadgets' and 'blogging', respectively) because of repetition and
> volume. I'm pondering a handicapping mechanism on over-used tags by
> one source. Something like "we'll only count 5 instances of any given
> tag from a blog per day". I think that will help the signal-to-noise
> ratio and boost the visibility of cross-domain tags more relevant.
> - More temporal clouds
> I'm also probably going to add in a weekly and monthly tag cloud view.
> This will provide stronger trending that is still relevant. More
> pondering required on this.
> - Inference: The holy grail?
> The next level of features I would like to add is related to inference.
> I would like to note connections between posts across blogs and time
> that may be related because of shared tagging vocabularies. Sort of a
> "6 degrees of kevin bacon" approach to content association. But
> that's a little out there at the moment and visualization is the long
> pole in the tent that I haven't figured out yet. Somebody suggested
> flash, but I'm not sure I want to go that route.