But one of the biggest problems with this tag cloud is to maintain the scale relationship. Imagine you have your “heart-shape” and try to fit in all the words, with which size do you start? Say you found a good, matching position for all but the last word: then you need to rescale everything to match properly.

The ImageChef works differently. In that approach its possible to use as much words as he wants. The advantage of this is that he can “fill” the heart shape until its filled enough (using smaller font every step). That is not just faster but also allows to create the shape.

A brute force method for this would be to just randomly check 40 times (just a number) for each size if a random position is free and positionate it there.

I’m trying to make word clouds that fill shapes (like http://www.imagechef.com/ic/word_mosaic/).

I’m looking to accomplish this either in actionscript or in illustrator javascript. Can you give me some pointers or help with source code?

Kind regards

Luke