Domain: puremango.co.uk
Stories and comments across the archive that link to puremango.co.uk.
Comments · 5
-
Re:A long-time problem
this is exactly how most session-based CAPTCHAs work. The timestamp idea is unworkable - it doesn't take that long for data to be ferried half way across the world, so if you implement a timeout, you'll end up pissing off your legitmate users as well thwarting spammers, and if you make the timeout longer it'll render it completely ineffective - what I'm saying is that it takes as long for a spammer to type a captcha as it does a legitmate user.
Stuff like "type this backwards in lower case" won't help *in the least* - it'd be trivial to get past, as trivial as writing a bot to collect email addresses, and we know how many of those there are.
Checking the IP address won't work (unfortunately) because certain ISPs (*cough*AOL*cough*) use multiple outgoing IPs for the same user; it's ridiculous but there you have it.
In any case, IP addresses can be forged; the spammer doesn't need to receive a response, he just needs to send his CAPTCHA and spam message; if he's on 4.3.2.1 and needs to send from 1.2.3.4 then he will - the server's "yes you got it" response will be sent to 1.2.3.4 but the spammer doesn't care; his spam has got through.
In short, there is no serverside way of preventing a captcha from being relayed to/from a 'processor' be it OCR or human.
However, what needs to be remembered is that in 95% of cases, any type of captcha will stop 100% of spam. Most captchas out there are pitifully weak in terms of OCR resistance, have implementation bugs coming out of their *ahem* and 'in principle' offer no security whatsoever, but they work because most spammers only after the low hanging fruit. -
Not just OCR
poorly designed captcha implementations can be circumvented 100% of the time, without having to use OCR. more info regarding this is available here http://puremango.co.uk/cm_breaking_captcha_115.ph
p (shameless self promotion - it's my site..)
also, it's no wonder that people are annoyed by CAPTCHAs - half the time they don't explain why the user has to enter the text, and almost all CAPTCHAs are developed around making the text hard to read. At the moment, it's only a few geeks who have managed to bulk-OCR CAPTCHA scripts. Generally even the presence of a totally insecure captcha is enough to stop spam dead in it's tracks - spammers just use a set script and fire it at a bunch of blogs, guestbooks etc; they are not currently targetting scripts at specific websites, and they're certainly not smart enough to perform bulk OCRing. -
Re:I don't use the Search Engine feature
yup yup yup. I currently use keywords for google, google images, traceroute, whois, ebay, wiki, xe.net, php.net, mysql.com (though their website is mostly useless (in comparision with the brilliantly useful php.net)), amazon, archive.org, a file extension search page, and ip2country. yay for bookmarks! your suggestions welcome.
FYI:
http://www.dnsstuff.com/tools/tracert.ch?ip=%25s
http://filext.com/detaillist.php?extdetail=%25s&Su bmit3=Go!
http://whois.webhosting.info/%25s
http://web.archive.org/archive_request_ng?collecti on=web&url=%25s
http://en.wikipedia.org/wiki/Special:Search?search =%25s
http://www.xe.com/ucc/convert.cgi?Amount=%25s&From =USD&To=GBP
http://www.ezwhois.net/index.php
http://search.ebay.co.uk/search/search.dll?satitle =%25s&ht=1&sokeywordredirect=&from=R8&fkr=1&soloct og=9
http://www.php.net/search.php
http://www.mysql.com/search/?q=%25s&charset=
http://puremango.co.uk/ip2country.php?ip=%25s
http://www.amazon.co.uk/exec/obidos/search-handle- form/026-9212734-6757257 -
Re:Simulating intelligence?If they can simulate something else than a virus (because I don't think viruses are intelligent) could they by this way obtain intelligence by simulating an intelligent animal?
From a philosophical perspective? Depends who you ask.
Most of the philosophers I know of are still talking about Turing's Computing, Machinery and Intelligence paper of 1950, which focuses on simulation of conversation (hence Turing Test) rather than learning systems or simulated life such as this.
JR Searle gives some pretty good reasons why, most notably in the 'Chinese Room' theory. From wiki:Searle describes a scenario in which a person is isolated in a room. The individual receives pieces of paper marked with Chinese characters from under the door. Even though the person does not understand Chinese, if there is a formal sorting process for the characters then they can be filed into a meaningful order.
Seale holds that the person simulating writing Chinese is not actually writing Chinese because they don't understand what they're doing.
In a similar way, a simulated intelligence wouldn't be self aware, and wouldn't have any intention behind it's actions.
On the other hand, people such as Turing would hold that it doesn't matter how the end result is acheived, if the end result is identical to that which it simulates, then there's no reason to suppose the machine is not intelligent.
With Searle's example though, I think it's noteworthy that with chat bots the analogy is fine, but with more advanced AI, the question of who designed the "formal sorting process" becomes crucial.
I wrote a paper on the subject of intelligent computers last year, IMHO it's not very good (it was the first essay I wrote for my degree), but if you're interested: http://puremango.co.uk/device.pdf -
id3 genre id limited anyway
the official standard (http://www.id3.org/) only defines 80 genres.
Nullsoft added a further 45 to the list, but these remain 'unofficial' additions.
you can see the whole list at http://puremango.co.uk/id3lib.txt