I currently use an approach I saw on Volker's site with a minor twist.
I have an image that called 1492.gif (for example) next to a field, we'll call "keycheckval", which asks for the visual numbers/letters from the image.
I have a hidden field called "keycheck" which stores the name of the image that I've loaded. (For this example, it'll store "1492.gif").
Now the user enters what they see for that image - and here's the trick - the image doesn't display "1492" but rather "1234".
Now, I post keycheckval ("1234") and keycheck ("1492.gif") to my comments agent which looks to a comparison table to see if the visual value of the paired image matches what the user entered.
1492.gif = "1234"
5512.gif = "4321"
blah.gif = "9999"
... etc,. etc., etc.!
So, if a user/bot enters "1492" instead of the of "1234" (thinking I paried the name of the image to it's correlating key value), the agent immediately returns false and won't process the comment. Otherwise, you either 1) are a bonified human-like person or 2) deserve to have your spam on my blog.
Hopefully that makes sense to someone other than me...
All I know is, that after implementing this counter I've yet to have a single SPAM entry in my comments.
I currently use an approach I saw on Volker's site with a minor twist.
I have an image that called 1492.gif (for example) next to a field, we'll call "keycheckval", which asks for the visual numbers/letters from the image.
I have a hidden field called "keycheck" which stores the name of the image that I've loaded. (For this example, it'll store "1492.gif").
Now the user enters what they see for that image - and here's the trick - the image doesn't display "1492" but rather "1234".
Now, I post keycheckval ("1234") and keycheck ("1492.gif") to my comments agent which looks to a comparison table to see if the visual value of the paired image matches what the user entered.
1492.gif = "1234"
5512.gif = "4321"
blah.gif = "9999"
... etc,. etc., etc.!
So, if a user/bot enters "1492" instead of the of "1234" (thinking I paried the name of the image to it's correlating key value), the agent immediately returns false and won't process the comment. Otherwise, you either 1) are a bonified human-like person or 2) deserve to have your spam on my blog.
Hopefully that makes sense to someone other than me...
All I know is, that after implementing this counter I've yet to have a single SPAM entry in my comments.
-C