As part of some investigation for my thesis, I made myself a little add-on for Mozilla Firefox that shows where in the page that JavaScript has been included. I'd been doing this sort of investigation by reading the code myself, but although that told me useful things, it wasn't ideal for communicating things to other people.
My add-on shows inclusion of new JavaScript (using a script tag) by putting a red border on the parent tag, and it shows JavaScript called from the onMouseover, onLoad, onClick, etc. attributes in blue.
One of the most interesting things I've found is that these are actually relatively predictable things. If there's an expanding menu, there's probably some JavaScript. Certain types of forms. Content that you'd expect to be external. Links that involve pop-ups. Embedded content from other sources.
Take a look at the way the add-on colours this weather site:
Once you've seen a few of the things it colours, you could guess a lot of the rest.
The question now is... Can this predictability be a helpful tool in developing more secure web pages?
Monday, September 15, 2008
Subscribe to:
Posts (Atom)