CAPTCHA has become a standard device to trap and block spam on sites like this one.
The idea is that if someone writes a computer program to drop spam comments on this blog, the computer program will be asked to solve a problem first. That problem (like “what letters are below in this squiggly image?” or “what’s 4+12?”) is not Turing-computible. So only a human being will solve it.
In the past 48 hours, I’ve had to remove two spam comments left by a human being. I can tell it was a human being because my site analysis showed a computer running Microsoft Internet Explorer 7, Windows XP, monitor resolution 1280-1024, located in China as the culprit. They solved the CAPTCHA first time. It must be a human. I’ve watched robots try and fail to solve my CAPTCHAs: They try 20 times in a row and then give up, and it never appears in my site analysis stats because no normal web browser ever loaded the page.
Time to try a new service. CAPTCHA is gone. In comes a new service called Mollom. Forget Tolkein. What this does is as follows: Someone leaves a comment. The contents of that comment are passed, encrypted, to Mollom’s servers. They check whether it’s “spam” or (wait for it): “ham”, and tell my website. I then block Spam, but let through all the Ham. Occasionally, Mollom returns “Not Sure” to my site, and then (and only then) you get a CAPTCHA to solve. See their how it works page for more information.
The upside: You can leave comments here without having to type in CAPTCHAs. Hurray. Especially as I know sometimes you patient visitors have to try a few times before answering the challenge successfully.
The upside: SPAM should be blocked automatically. If the SPAM-masters pay a human being to type it in, it will still be blocked.
The downside: There’s a privacy implication. Any comment you leave has to go to Mollom for verification. If this concerns, you, you’d better see their Privacy Policy
Now all I need is lots of Ham please to test the thing. If you have problems, let me know.
Comments
Testing this out...
OK: I had to test.
I did not get asked for a CAPTCA, and my comment was accepted.
The real test is whether or not SPAM starts to get through.
I didn’t understand a word
I didn't understand a word of that.
Not feeling smug, Neil...
Topic: I'm talking about the fact that you no longer have to decipher a picture containing some obscured letter and numbers before you can comment.
Focus: Reasons why - (i) spam gets through, (ii) it's inconvenient for genuine commenters, (iii) there's a better alternative
But I like ham
I've run a promotion on Edible Words on my blog this morning. I wonder if Mollom likes adverts like this one.
Morning Neil
Mollom didn't seem to mind the promotion, as it let you through fine. It's yet to block any real spam though, so the jury's still out as to whether it's any good as a gatekeeper.
Once it's been thoroughly tested here, I'll eventually remove any CAPTCHAs there may be at Edible Words.
Why isn’t 4+12
Why isn't 4+12 Turing-compatible? I should have thought that was exactly the sort of problem computers are excellent at.
Math Captcha
Yes - I'd have thought that what is called a "math captcha" is an oxymoron.
The image ones are best, I think - especially if coupled with the option to read it out for visually impaired users. But I hit problems once spammers started paying people in China to answer the CAPTCHAs for them. I suspect mathematic ones would be more of a deterrent, because they are non-standard - but they are more prone to being solved by robots.
So I gave up and went for Mollom - which has (so far) let nothing through it shouldn't and blocked nothing it should have let through.
Add new comment