124Views34Replies

Author Options:

Bot is Using An Account to Post Spam Answered

Admins, please help instructables by deleting this user: https://www.instructables.com/member/ghuif87dsa9/?show=comments&limit=500

The account was created today and has already posted the following comment on 25 instructables:

"Your satisfaction is our eternal pursuit.

( http://asso.in/url/3qy )

We specialized in the exportation of sport shoes and other products(clothing, bag,sunglasses,watches,belts,etc )which have great enjoyed popularty in the world market Many of our goods are on sales ,we can guarantee the crediblity by Pay-pal and delivery time .we would like to make a long termship.

1)Name : The perfect gift.
2)Grade : AAA+
3)Package : in original boxes
4)Color : various
5)MOQ : 1 pc
6)Payment : Pay-pal/Western Union / Credit card/Moneygram
7)Ship-ping : 4-7 days with guarantee of customs clearance, drop shipp-ing is accepted to customers'demands.
8)Who-lesale: very low price, you can make a small order first to test quality and service.

( http://asso.in/url/3qy )"

Tags:spambot

Discussions

0
None
Kiteman

7 years ago

Thanks for reporting this!

0
None
StumpChunkmanKiteman

Reply 7 years ago

They're actually quite the evil spam bot. Their ip address constantly changes, they make several accounts a day, each one posting right around 25 comments then moving on to a new account, and they keep changing stuff around on us...

...we'll get 'em though...

...it just might take a little time to figure out their trickeries!

0
None
RedneckEngineerStumpChunkman

Reply 7 years ago

Just jumping in the middle of this conversation but why not have a posting limit for the first 48 hours on new accounts. This would limit the amount of Spam and might alert the admins faster to these accounts. Or it could be just like submitting a ible and the poster could see the comment but its not seen till 48hrs later if it's not flagged by the filter or admins. Just a thought.

0
None
caitlinsdadRedneckEngineer

Reply 7 years ago

That would ruin the magic for the thousands of paying public people that walk through those doors every day. People have an expectation and want instant gratification... Walt D.

0
None
KitemanStumpChunkman

Reply 7 years ago

Track down their physical address, and post them a package of neodymium magnets...

0
None
StumpChunkmanKiteman

Reply 7 years ago

hahahahahahahaha...Not a bad idea!

After I posted that last comment, I reread the post and saw "Ship-ping" and "Who-lesale" and started laughing. They know that their stuff is getting caught in filters and their actively changing the things they think might be culprits.

In the end, all this will do is give us much more accurate spam filters.

0
None
KitemanStumpChunkman

Reply 7 years ago

Maybe put comments through a spell-checking filter? Too many odd words, and the filter is tripped.

That would stop a lot of the poorer Instructables hitting the lists as well...

0
None
StumpChunkmanKiteman

Reply 7 years ago

Hah! Nice! Just disallow bad grammar completely from the site. If you can't spell or use close to proper grammar, we don't want your comment.

(To be fair, I'd probably be one of the ones who's posts constantly got caught in those filters.)

0
None
KitemanStumpChunkman

Reply 7 years ago

OK, maybe misspellings of a certain proportion of words over a certain length?

Frequency of hyphens or dollar signs?

I'm actually trying to help here...

0
None
StumpChunkmanKiteman

Reply 7 years ago

We're actually working on implementing new code that should be a nice catch all...we'll see how it works.

0
None
kelseymhStumpChunkman

Reply 7 years ago

Just a question. How come you guys are trying to do the arms race manually, rather than deploying a third-party (or even roll your own) Bayesian selector? You could tie the selector to the use of of the "Spam" flag by Admins (to reduce false positives) and just let it learn.

If Admins can flag user accounts directly, the selector could even learn (in principle) features of spambot usernames or profile fill-ins.

0
None
StumpChunkmankelseymh

Reply 7 years ago

Update! Just had a chat with Rachel. She said the reason we don't have a Bayesian selector, is because no one hear has ever thought of writing one. She also pointed out that she hasn't written one yet, and that from what she knows of it, it sounds like something fun to write. She also wanted me to thank you, and let you know this is why we love you. ^_^

0
None
kelseymhStumpChunkman

Reply 7 years ago

:-) No problem! Bayesian statistics is just another little gift from the physicists (okay, we stole it from the mathematicians, but whatever).

0
None
steveastroukkelseymh

Reply 7 years ago

Bayes was a minister ! I don't know if he was a mathematician, moonlighting or a minister, moonlighting though....

0
None
steveastrouksteveastrouk

Reply 7 years ago

You had to be a Doctor of divinity I think to graduate from Oxford or Cambridge.

Steve

0
None
kelseymhsteveastrouk

Reply 7 years ago

Darwin was also a minister. I think participation in the C of E was a requirement for higher academia in the 19th Century.

0
None
JayefuuStumpChunkman

Reply 7 years ago

Wouldn't it be better to use one that's already written and fix a few more of the thousand bugs that's been backing up with the time saved? Like the no line breaks in PMs, that's particularly annoying.

0
None
GoodhartJayefuu

Reply 7 years ago

you can use html in the pm's so you can place a < br > for a line break :-)

0
None
StumpChunkmanJayefuu

Reply 7 years ago

She only expressed that it sounded like fun to write one, since she hasn't, but I have no idea how she would implement it. To be fair, before it gets to that stage there probably has to be several dev meetings talking about what exactly we need and how exactly to go about achieving it. I realize bugs suck, but sometimes it's much harder to figure out what's causing a bug, or exactly where in the line of code to fix it, then it is to just put something new in there. But like I said, it'll only be done when they have sufficient time to do it, so no worries.

0
None
StumpChunkmankelseymh

Reply 7 years ago

That sounds cool! I've never even heard of a Bayesian selector, let alone know how to implement one. I'm sure other, much smarter people working here do though. How would it work? Would it bypass people completely and be it's own kind of bot, on a search and destroy mission for spam on the site? Or would it just troll the site looking for spam, and put it all in an area for people to look at later? Oh! Or does it do both?! it sounds cool...I'll have to bring it up to Rachel and see what she knows about it.

I do know that we like having a hand in things. I think that if we took that away, while things might go a little faster, we'd lose a little bit of touch with what's going on.

0
None
kelseymhStumpChunkman

Reply 7 years ago

You could write your own (and I am sure that someone like Rachel would have a great time doing so).

But there are lots of commercial products out there that can do it for you, in a flexible, configurable way. You don't have to lose all control -- the selectors work best if they are actively trained and monitored, so they don't go off trapping otherwise perfectly ordinary posts (cough :-).

In fact, if you're using Outlook, you're already using Bayesian filtering on your e-mail. When you flag something as "Junk," (or when you grab something out of Junk and tell Outlook how stupid it was :-) that gets fed back into the filter to adjust the probability functions, so it does a marginally better job next time.

0
None
steveastroukkelseymh

Reply 7 years ago

+1. Software writers LOVE re-inventing the wheel. BUY IT IN. OR USE AN OPEN SOURCE TOOL.

Steve

0
None
caitlinsdadsteveastrouk

Reply 7 years ago

Good software writers/programmers know how to take a wheel and put it on alloy rims or retread the tire.

Mediocre software writers/programmers know how to take a wheel and patch it up and maybe tighten up a lug nut or two.

Lazy software writers/programmers just mount the wheel and hope it doesn't fall off.

0
None
Lithium Raincaitlinsdad

Reply 7 years ago

If a person/team is dealing with a ton of SPAM and it has never occurred to them (nor even come up in their research) to write a Bayesian filter, that person/team is very probably not equipped to make any new and wonderful innovations that justify reinventing the wheel.

0
None
caitlinsdadLithium Rain

Reply 7 years ago

So what you are saying is: Call AAA to come with the tow truck since they have the equipment and expertise to change the tire?

0
None
Lithium Raincaitlinsdad

Reply 7 years ago

Yeah - you're welcome to try and change it yourself using that rusty spoon and garbage bag tie, but I daresay it will go better if you just call AAA. :p

0
None
caitlinsdadStumpChunkman

Reply 7 years ago

I know it would not be politically correct, but a little profiling would catch a lot of those spammers from the git-go. They seem to be signing up account with certain forms of usernames and the gender is stated to be female. The avatar picture selected by most of them is also a clue to spot them easily.

0
None
kelseymhcaitlinsdad

Reply 7 years ago

Hey! How come your reply got through, but mine is still stuck in the keyword filters?!?

0
None
Goodhartcaitlinsdad

Reply 7 years ago

Well just so they don't start breaking symetry.....

0
None
Lithium Raincaitlinsdad

Reply 7 years ago

Just what are you trying to say?! You got something against certain forms of username, stated female gender and an avatar picture? Huh? DO YOU, punk?!