Our website would like to use cookies to store information on your computer. You may delete and block all cookies from this site, but parts of the site will not work as a result. Find out more about how we use cookies.

Login or Register

Powered by
Powered by Novacaster
 
'Kittenauth' verification system
by Bruce Ure at 12:14 27/07/06 (Blogs::Bruce)
Now this is cool. (If link breaks try kittenauth.com).
It is designed to separate humans from computers, being a variant of the distorted-font CAPTCHA system.

It presents you with a grid of pictures of animals and you pass authorisation if you can successfully click all the kittens (or pandas, or foals, or whatever animal it tells you to click) from amongst the other species.

It works extremely well. Much simpler and quicker than text-based CAPTCHAs, and you won't run into the reasonably frequent ambiguity of the is-that-a-4-or-an-h? type.

It was only a matter of time, I suppose, before it got adapted. Someone came up with the brainwave of using images from amihotornot.com (which has an API, can you believe it) and the test becoming to click the 'hot' folk and ignore the not-hot.

A perhaps not terribly respectful adaptation, but if you consider that the scores on amihotornot reflect the collective opinion of often thousands of folk, likely to be very accurate and reliable.

You'd think.

Maybe I have weird taste, because I find it much more difficult to spot the 'hotties' than to spot the kittens.

I came across this stuff while looking (again) into ways of preventing comment spam on my wordpress blog. I think I may have a go at implementing kittenauth.

--

<< Just a quick note of... Excellent error messages >>
View Comments (Threaded Mode) Printer Version
'Kittenauth' verification system Bruce Ure - 12:14 27/07/06
Re: 'Kittenauth' verification system Jonathan Tuppeny - 15:30 27/07/06
It's so sad that so much effort has to be exerted to bypass the bloody spam.

I read an article on the BBC site today that said that only 3-4% of email received was actually genuine mail.

I would suggest that something should be done but let's face it the spammers will probably find a way around anything. It's depressing.

--
jht

Re: 'Kittenauth' verification system Bruce Ure - 15:46 27/07/06
They will find it very hard to get round this. They would have to either exploit some weakness in the system (which once reasonably mature would be nigh-on impossible), or devlop a machine that could tell the difference between the animals (completely impossible at present, I'd guess). Or brute force it, but you code against that.

Sadly though, the one way round it that is likely to be used is to pay people in poor countries a pittance to sit there all day and click kittens.

--

Re: 'Kittenauth' verification system Jonathan Tuppeny - 16:02 27/07/06
Don't get me wrong I thing this is great method for blocking unwanted entries but I suspect that they'll come up with something that will get some stuff through - maybe just 1 in 100 attempts.

I'm sure the first spam filtering software writers thought they'd cracked it :-(
--
jht

Re: 'Kittenauth' verification system Simon - 16:13 27/07/06
pay people in poor countries

Nail well and truly hit on head.

Humans are the cheapest computers on the planet, but I suggest that increasing the perceived value of a human life is likely to be a harder task than defeating spam.
--
simon

Re: 'Kittenauth' verification system Simon - 16:09 27/07/06
Email's dead, it's just not stopped writhing yet.

The way forward (pontificateth he) is a mechanism where you don't send an email anywhere, you just put it on your own message server and tag it as being for one or more intended recipients. Other people do the same on their message servers.

People who you've previously approved then periodically poll your message server asking if there's anything for them, and your message server periodically polls the message servers of your past correspondents seeing if there's anything for you.

Finally, instead of all 'email' being delivered into a particular recipient's mailbox, this new 'not email' is put into a box on your machine tagged as being "From" that sender. So you have lots of "Sender" drop boxes instead of one "Recipient" dropbox.

You can then erect whatever barriers you like to becoming one of the lucky few granted permission to poll and deliver to your local message server, and that keeps the spammers from reaching you.

As for someone you've never heard from before legitimately needing to get in touch, well maybe you don't erect such quite high barriers for permission to put stuff in the drop box for 'new correspondent requests', and you continue to filter that for spam leaving you to concentrate on the stuff you know has come from legitimate senders via your pre-approved channels.

--
simon

Re: 'Kittenauth' verification system Jonathan Tuppeny - 16:24 27/07/06
I do get new business enquiries via e-mail so I'd be loathed to erect any barriers that might stop these.

I also don't feel comfortable about surrendering email to spammers but I think you may have a point - fighting spam could end up being a pyhric (have I spelt that right?) victory.

I think this actually just echoes my original point about how depressing it is that people should have to expend all this effort to escape brainless marketing.
--
jht

Re: 'Kittenauth' verification system Steve - 21:17 27/07/06
One idea I have that might stop spam is to use a device called a parchment which acts similar to the Whiteboard in Messenger. You'll need to obtain another device capable of recording signals on the parchment and some skills is likely needed to reproduce the gylphs from the standard ASCII character set on the parchment but I've seen studies that suggest that most users pick up the skill fairly quickly.

The parchment concept divides the message body from the transport method and allows them to be configured independently. A common transport device is similar in style to a parchment and has the same inherent signal absorbing qualities but it has been folded in a way that creates an internal space for insertion of the original parchment. As an analog, consider the solid casing in which a typical Xbox game is stored but a): thinner, b): more flexible and c): usually white instead of green.

The destination ID of the transport is typically recorded on the front of the transport parchment (technically known as an "envelope" after the TCP/IP envelope). It is usually recorded using the same signals used to record the users message onto the enclosed parchment but it may also be recorded using the little known Mail Wizard feature of Microsoft Word 2014 by following steps 1 through to 237 in the user manual.

Delivery of the content is accomplished by manually inserting the "envelope" (taking care to ensure that the enclosed recording is properly sealed) into the CD slot of the red transmission devices to be found installed on most high streets thanks to the assistance of a technology grant by Charles v2.0. A token is required to be attached to the envelope which can be purchased separately which is required to access the service. The transmission protocol utilised by these devices is fairly slow but mostly reliable. A sophisticated error correcting protocol ensures that the majority of envelopes arrive at their final destination even if the destination ID is not sufficiently complete. In some cases and error packet is returned back to the originating service with additional data appended to indicate the cause of the failure (frequently E_ADDRESS_NOT_FOUND).

Owing to the slow transmit rate and the additional cost involved with this service, there is less business justification for utilising it for sending spam although the opportunity exists. However compared to the 30 year old e-mail technology, the ratio of spam to valid content is typically a positive ratio.

--
stevepa

Re: 'Kittenauth' verification system Simon - 09:22 28/07/06
It'll never catch on, you know.
--
simon
Re: 'Kittenauth' verification system Gordon Hundley - 02:21 29/07/06
I think that one way forward to reduce spam would be to marry non-repudiation with an accrediation system that uses reputable authorities. As happens, you look up these thoughts with a search engine and discover that people are way ahead of you, and have expounded such ideas more coherently in the public domain:

http://www.openspf.org/aspen.html
--
DrGoon