Author: P.Karn (26 Jan 05 5:17am)
A popular technique to thwart machine harvesting of email addresses or the automated signing up to accounts, mailing lists, etc, is to give the user the information in an image of some text, usually with the characters mildly warped in some way that humans can easily read but OCR programs cannot.
It is often pointed out that this technique is unfriendly to those who cannot easily see an image, especially the blind who rely entirely on text-to-speech synthesizers.
So how about supplying an optional audio link that speaks the email address or token as an alternative? Seems pretty obvious to me. I know that speech recognition is getting a lot better, but it still takes a lot of CPU power and still isn't as good as a human ear/brain. If you can't record the audio clip in your own voice, use a text-to-speech synthesizer.
Comments?
|