Can simulated handwriting stop the spambots from getting through?(technologyreview.com)
technologyreview.com
Can simulated handwriting stop the spambots from getting through?
http://www.technologyreview.com/web/21939/?a=f
3 comments
If that works --- it will create a lot of jobs in China and India.
I can't read any of the sample CAPTCHAs provided.
Something I have not seen pointed out yet (in articles and discussions over the preceding many years) is the fact that CAPTCHA tests are essentially Gestalt Completion tests. A Gestalt Completion test is a type of IQ test. As CAPTCHA tests get progressively more difficult to solve, they might weed out progressively larger portions of the population - the human population, that is.
Perhaps, as a side-effect, the quality of web-based discussion will rise.
.
From Arthur Jensen's 1980 book Bias in Mental Testing (pp157-158):
Gestalt Completion. Familiar words and objects are shown in various degrees of "mutilation" and the subject must infer what the word or object is. Performance is scored according to the average time taken per item or the number of items completed within a given time limit. The difficulty of any given item is a function of its familiarity and the extent of the "mutilation"; these can be varied independently in making up items. ... Gestalt completion items, which are known to be a good measure of the general intelligence factor (as well as of a small group factor labeled speed of closure), merit further experimental investigation in their own right.
Perhaps, as a side-effect, the quality of web-based discussion will rise.
.
From Arthur Jensen's 1980 book Bias in Mental Testing (pp157-158):
Gestalt Completion. Familiar words and objects are shown in various degrees of "mutilation" and the subject must infer what the word or object is. Performance is scored according to the average time taken per item or the number of items completed within a given time limit. The difficulty of any given item is a function of its familiarity and the extent of the "mutilation"; these can be varied independently in making up items. ... Gestalt completion items, which are known to be a good measure of the general intelligence factor (as well as of a small group factor labeled speed of closure), merit further experimental investigation in their own right.