discuss: Thread: recaptcha

Do you know this, and is it worth help?

http://recaptcha.net/

jdd
-- 
jdd for the Linux Documentation Project
http://wiki.tldp.org
http://www.dodin.net
http://www.facebook.com/profile.php?id=1412160445

Quoting Jean-Daniel Dodin ####@####.####

> Do you know this, and is it worth help?
> http://recaptcha.net/

Wow, that has to be about the most clever piece of social engineering
I've seen in years.

-- 
Cheers,              Crypto lets someone say "Hi! I absolutely definitely have 
Rick Moen            a name somewhat like the name of a large familiar 
####@####.####  organization, and I'd like to steal your data!" and lots 
McQ!  (4x80)         of users will say "OK, fine, whatever."    -- John Levine 

Maybe I don't understand reCaptcha, but it looks to me like it would
make enough of a web page unreadable enough to give you a migraine.
And what if you try downloading a page say, onto a Palm or a phone or
a text browser?  Will the page render the pixelated text correctly?

I think the result is similar to printing a page as a PDF then
converting it to a JPEG then posting that to your site instead of the
original text.  If so, I think it carries all of the same problems of
not being able to select, search, enlarge or comfortably read what
you've written.

But, again, I could be misunderstanding it.

Borden

####@####.#### a écrit :

> But, again, I could be misunderstanding it.

I think you did

recaptcha is only a captcha like the one we use already on the wiki,
but asking for two words, not one.

one of the two is a word from a true book, unreadable by OCR, so the
user make a small contribution to the book digitalization giving his
answer

the other is the real captcha

jdd


-- 
jdd for the Linux Documentation Project
http://wiki.tldp.org
http://www.dodin.net
http://www.facebook.com/profile.php?id=1412160445

I'm still confused on how it works, but thank you.  It probably does
need changing since the bots will figure out the answer to all the
current wiki captchas is 'Paul Jones' pretty soon!

Borden

Quoting ####@####.#### ####@####.####

> I'm still confused on how it works, but thank you.  It probably does
> need changing since the bots will figure out the answer to all the
> current wiki captchas is 'Paul Jones' pretty soon!

It would be really, really rare for anyone to design/maintain a
_site-specific_ spambot for just one target site.  What exists out there
are generic, site-independent scripts that roam the Internet looking for
any Web forms whatsoever, and blitz them with submissions.  (It get an
insane amount of submissions to the links feature of my PerlHoot-based
knowledgebase at http://linuxmafia.com/kb/ , and none of that ever makes
it to public display because all submissions are moderated, but that
make no difference, there being no intelligence whatsoever behind the
spambots.)

So, in general, having a site-specific "answer this trivial question"
feature is more than sufficient protection -- because the spam is
overwhelmingly fully automated.