Man or Machine: Who Can Find Your Inbox?
Feb 16, 2010 by Aaron Rubman
For the life of the Internet, the art of protecting online e-mail addresses has rested upon the ability of web developers to come up with effective ways to sort man from machine. In short, a CAPTCHA.
If a machine can identify and grab your e-mail address, you will become a target for SPAM - but removing your e-mail altogether is not an option if you want to provide legitimate clients and leads with a way to reach you without having to leave their computer.
When I first started looking at website code (back when a 28.8 baud audio modem was state of the art) you could fool machines by replacing each letter with a specific three digit code. At the time I wondered why the code switched from plain text to an indecipherable block of numbers whenever I reached an e-mail address.
Eventually I realized that each of those numbers corresponded to a letter - and so did everyone who wrote the programs on the prowl for e-mail addresses. With a simple addition to their code, these e-mail harvesters were once again on the prowl.
Code alone, it would seem, was insufficient to separate human visitors from machine.
In response to this development, another solution came to popularity - placing e-mail addresses within graphics (and later within flash animations).
At first the only problem with this approach was that you could not link these pictures directly to an e-mail address (as the process of doing so would once again make the address detectable within the code of the page).
However, as optical character recognition (OCR) has grown more sophisticated, these methods have become less secure.
On the whole, OCR is quite useful. Any time you scan a document directly into a word processing or spreadsheet program you are using some form of OCR. However, because OCR has become so good - machines can now just look at how a page is presented to visitors (instead of at its code) to determine if any e-mail addresses appear.
However, just as OCR posed a problem, its use also presented a solution. Every so often OCR programs will misread words. These words, then, can be used to sort a human from an OCR using machine. ReCAPTCHA offered a service that could be integrated into form and forum protections.
MB/I can now place just such a protection on your e-mail addresses, making them harder for spammers to gather and exploit.
While no protection is absolute, requiring human input goes a long way towards keeping your e-mail address out of the hands of abusive e-mailers.














Recent Comments