pouët.net

It seems since 21 days ago (as of writing this) that there's a person (persons, or bot) creating a bunch of users with a red flame on a black background avatar, no other details, and a presumably randomly generated username. So far most seem like they haven't done anything at all other than exist.

What's up with that?

start here and hit "next page" a bunch

m77 back to have a rant and get quickly b&?

Seems to be going on for months already, haha.
Maybe someone thinks pouet-elections are soonish, trying to generate an absolute majority ;)

The strangest thing about them is that they don't do anything.
And well, they seem to use only 2 avatars: the flame one () and a shrunek pouet.net screenshot one ()

Maybe a good solution would be banning all the users with these avatars who don't have any glops. Though there's a risk of an innocent lurker getting caught by it, in which case it would be nice to put up a contact form through which one could ask a mod to get unbanned.

They can't post because they're unable to activate their accounts.

Oh! then let's wipe the accounts that haven't been activated within a week or so. No harm really.

Could be someones userscript maybe, doing strange things tho. Like getting access to some data by taking the first hurdle thru generating a one-time-used user...no real idea, just sth like this could be whats causing it.
But the glöperators seem to be aware of it and just kept shut until now...to let the trap hit at the right moment maybe. In this case, this thread didnt help!
It´s going on for more than 21 days btw, also there are more than 2 avatars used with new accounts that have random names. Just go back some pages more, i kept seeing them 2 months+ back, then stopped following the backtrace, had to go babysitting.

what p01 said. also add captcha or even a better captcha (dunno if there is one atm) to the user registration form maybe.

just a 'rand(1,5)+rand(1,5)='-to-check-youre-a-human-captcha would do the trick i guess

anti-captcha would be to put some math into the bot-code then. unless the captcha would be in form of an image. while image-deciphering anti-captchas aren´t new aswell...just they don´t work that good in most cases, hehe.

Math captchas are everywhere and thus easy to solve automatically by bots. Best thing is still to use custom questions that are specific to this site but still easy enough to solve (e.g. not stuff like "what's the colour of the sun", but rather "is this a site about the demoscene or the emoscene?", "spell pouet backwards"). You can even go as far as putting the actual answer on the site, as long as there's not a human being that's solving the questions, the bots won't be able to solve it (unless someone specifically trains them to register on scene.org).

Also hardy, please stop bullshitting about things you obviously only heard about but never dealt with yourself, pretty much all image captchas are pretty week... Google tries to keep with the spammers but in the end that only makes reCaptcha harder to solve for humans.

*waiting for someone to blame razor again*

pretty sure it's razor guys.

Saga: Who says i only heard about it? :D
You said it yourself: it´s getting harder (almost too hard by now, i am _guessing_ like 2 of 5 correctly here) for humans to solve them images...while the coders of anti-captchas for images almost all gave up by now.
(desaturating, resaturating, let the algo check for probable solutions, try this, try that, check again, start over from original image, begin with this or that, check, then desaturate, resaturate, check, still nothing, go back to original image...and so on. -> even intelligent algos dont get the most of em by now! :P)
I am a scener/coder for 25+ years by now...so please dont tell me what i obviously only read about!
*stillinagoodmoodhere* (just wanted to answer on your try to make me feel dumb!)

I am aware you´re trying to say that image-captchas are weak and easily cracked, also i am aware i am not the coder to be asked about webCoding-stuff, but i am just contraire to that, as i see it from a different angle. Maybe image-recognizing-software was good the last years, but _by now_ (as often used in my last post) they just have no chance against the quality of the captchas in most cases anymore. same as with humans.

maybe adok can do a discount iq test consisting of lets say 5 questions on logic and eugenics, that'll teach them bots!

random question: what's wrong with the number system of users. there a couple empty slots? how does that work?

and one can find gargaj the po(u)et. :D isn't that his language?

y0 hardy, and there is still the good old images of dogs and cats thing... put a bunch of images, ask in which there is a dog or a cat. last time i read about it it was pretty much unsolvable. you need some rather large collection though.

Maybe in form of a riddle? i made one, as adok is _slowly_ fading away from the scene anyways, atleast according to his own words, he´ll only be there for another DAMN 3 months and 2 weeks!

Adok´s Riddle:
1.) who´s apparently not ze boss of ze mensa?
2.) who´d like to be ze padawan of ze boss of ze mensa?
3.) who thinks he´s one of ze chosen ones?
4.) who also thinks he´s a good coder?
5.) who releases stuff no1 likes?

This site:

O is your online demoscene resource
O is for posting ponies and unicorns
O is your secret kitty pic stash

Radio buttons don't work (?)

This site:

O is your online demoscene resource
O is for posting ponies and unicorns
X is your secret kitty pic stash

they do, apparently

multiple choice, yeah!
8 possible solutions in this case! ofcoz only all 3 are logically true!

psonice: gratz on making a working captcha!
(which is only not working for the first 7 attempts the bot-creator/coder does! from there the answer is known and implemented in the bot, lost!)