* the word that was double-clicked upon
* the start offset of the word from the beginning of the div (meaning the number of characters of text between the beginning of the div and the beginning of the word).
* the top, left, and height of the selected word, in pixels.
Note that these requirements are trickier than they might sound at first. For example, if the div contains the html "create a <i>super</i>conductor for the experiment" and the user double-clicks on the word "superconductor", then the word will span across multiple nodes. The function must work equally well on Mozilla and IE6/7. Note that in Mozilla, there is no easy way to obtain the offsetTop and offsetLeft of a Range. If you resort to inserting an element to determine these, your code must also remove the element, leaving all document nodes to form the same exact graph they initially had.
You must fulfill the requirements exactly. If there is some edge case that your code does not cover, then it is not really a solution. I would suggest that you not accept the task unless you are *really* familiar with the DOM implementations in FF and IE. The code must also be well-designed, meaning (for example) it can't be sloppy and method bodies must be under 100 lines.