Spam Protection Tool

The JavaScript code generated here performs several non-trivial computational tasks to protect addresses against spam. This contrasts with the simple entity code replacement or string concatenation done by most anti-spam encoders. To be able to perform the computation prescribed in the script, a software application must be able to understand JavaScipt. All recent web browsers do. This is however not the case of webcrawlers. Why not?

• Role mismatch  The role of crawlers/robots and other spambots is to gather a large amount of data from web documents. They are non-interactive programs capable of running without human intervention for days. At the other extreme, the role of JavaScript is to provide some interactivity in websites.
• Inefficiency  If someone wanted to construct a crawler capable of interpreting JavaScript, he would most probably reuse the JavaScript engine of a browser, which is do-able. Still, would it be worth running all the code available on the web to extract a few email addresses, knowing that only a tiny proportion of it is used for spam protection? Clearly, No.

A limitation of the previous approach is that it relies on a valid and complete program: given an appropriate tool, it can be executed and reveal your email address. We can resolve this by making the program computationally incomplete with regards to the information we are trying to hide. The program now does not contain your full email address, only part of it.
However, since we would (obviously) like legitimates users of your website to be able to contact you, we still need to store the missing information somewhere. Somewhere readily accessible to them, but not to crawlers. Here we store this information on a picture. Reading it on the picture is reasonably straightforward for humans, but it is a problem very hard to solve computationally.