Font Weighting Not Working?

Actually, after messing with it some more, it is better than it was, but the font weighting is still not near as accurate as it used to be.

Additionally, the "randomness" of re-visualizing is still not very random. It only changes two or three words and gives you two or three different options instead of really mixing things up like it used to.

Thanks for the work... it is appreciated. I'm just giving my feedback.

You are right regarding font weighting. It is not that accurate like it was before. It is the drawback of new algorithm. But new algorithm provides much better shape fitting that I consider worth of sacrificing some font weighting.

"Randomness" is really dependant on how many tag rotation angels you choose. E.g. if you chose only 1 rotation angle then there will be no randomness at all. The more you choose the more random your cloud will be when you visualize it.

I agree with mattuck and rolechowski that this is the one major drawback to the api. I love this design, but there are certain clouds where one needs accuracy. Shape is great to convey generalities, but when using the cloud to actually convey the exact information/weighting on particular words, the size is critical.