Recognizing forum spammers

[Someone]

@LowLevel

In most cases, it’s an attempt to create a link to another website, mainly because search engines use the quantity and quality of links to estimate the popularity and authoritativeness of websites.

From my own experiences I wouldn’t say “mainly”: Most spam links I had to kill in the past referred to sites where you can buy drugs (mainly the one for men…) or sites with malware.

@milanfahrnholz

In that case I´d like to suggest to @RonGilbert to pin this topic.

I second this.

[LowLevel]

@Someone

From my own experiences I wouldn’t say “mainly”: Most spam links I had to kill in the past referred to sites where you can buy drugs (mainly the one for men…) or sites with malware.

You’re right. That’s a big part of the reasons why links are created. I tend to observe more phenomena about links created for marketing purposes, but I’m not sure if it’s actually the biggest cause for link spamming.

[Sushi]

@LowLevel

unless someone will invent some kind of captcha not solvable by captcha-solving services (which is really a tough goal).

I think the “seckrit question” on the blog was pretty effective and unsolvable most of the times

[LowLevel]

@Sushi

I think the “seckrit question” on the blog was pretty effective and unsolvable most of the times

CAPTCHA solving services use real humans in less developed countries who are paid a few cents to do whatever a normal user needs to do to pass any kind of check.

Also, I’m pretty sure that there are software applications that can solve math puzzles even when they are expressed in natural language instead of math formulas. There was a famous one that had 99% accuracy but I have just searched for it and it seems it doesn’t exist anymore. I bet there are several similar applications, though.

[RonGilbert]

@Sushi

I think the “seckrit question” on the blog was pretty effective

It was. In the 2+ years the blog was active, I deleted 5 spam messages. The math problem caught 22K attempts. All the spam is still in the DB, it just doesn’t get displayed. It works because it’s a custom solution on a very small site. If they did the same thing here, it would be worth spammers time to get around it.

[Guga]

@Sushi

I think the “seckrit question” on the blog was pretty effective

So effective, even I had problems in solving it sometimes they say it was because it timed off, but I prefer to think that it’s because I’m a bot.

[Sushi]

That’s exactly why it was so effective. Now, what could be a better way to discourage spammers is a reverse timer thing. So during the first 15 minutes you’d always get the question wrong and only afterwards it would accept the answer. Because we, real followers of the blog usually spend more time writing our posts and spammers don’t. They just copy and paste a generic line and want to move on. Keeping them busy for fifteen minutes failing to solve real easy math is just karma.

[ZakPhoenixMcKracken]

@Guga

So effective, even I had problems in solving it sometimes

With @Nor_Treblig’s script, even this task was automated!

[Nor_Treblig]

@LowLevel

I am newbie here. I would like to say hello to all of you guys.

Hi, welcome to the forums!
I think this piece of software may help you solve your problem: click me
I don’t know this guy but It’s only $9.99 and since I’ve got it I have never looked back, it’s working like a charm!

---

@LowLevel

I’m pretty sure that there are software applications that can solve math puzzles

If you want to see my advanced math solver (which isn’t even using binary trees) take a look here and search for seckrit_question
It could even solve advanced mathematical problems (for Rons and Nors that is), like subtraction

@LowLevel

using captcha-solving services

Captcha-solving services are really great because captchas got just too hard to be solved by myself!

[PiecesOfKate]

@Nor_Treblig

Captcha-solving services are really great because captchas got just too hard to be solved by myself!

Agreed.

‘Select all the images with store fronts.’

• squints at screen *

&#£@%*!

Edit: * Discourse turns asterisk to bullet &#£@%*! *

[Guga]

@PiecesOfKate

Edit: * Discourse turns asterisk to bullet &#£@%*! *

You have to escape your asterisks
Write \* and you’ll be fine.

*squints at screen*

Edit: note that I had to escape that backslash too, so to write \* I actually wrote \\*, and to write this \\* I had to write \\\\\*, and to wr… well, I think you all understood.

[LowLevel]

@Guga

Edit: note that I had to escape that backslash too, so to write * I actually wrote \, and to write this \ I had to write \\*, and to wr… well, I think you all understood.

There is another option to avoid writing all those backslahes when you want to cite MarkDown syntax (and also to prevent MarkDown from making a mess when people will quote your text, as I did above): write it as preformatted/code text, which shows everything literally:

In this example I wrote exactly \* and it will be shown just as \*

You can also insert preformatted text inside any normal sentence, like this, but it needs the backtick character, which is not easily accessible on my keyboard, so I never use it.

Do you think that posting in this section a short cheatsheet of the basic MarkDown syntax would be helpful to others?

[Guga]

@LowLevel

write it as preformatted/code text, which shows everything literally:

In this example I wrote exactly * and it will be shown just as *

You can also insert preformatted text inside any normal sentence, like this, but it needs the backtick character, which is not easily accessible on my keyboard, so I never use it.

It doesn’t seem to make a difference in the quote, the backslashes disappear anyway.

The best thing would be if Discourse, instead of taking the displayed text when clicking quote, took the source of said text. The software can already perfectly see from where the quoted text comes from - author, post, topic - I don’t think it would be too difficult to extract the internal text from the same post.

[Nor_Treblig]

@Guga

It doesn’t seem to make a difference in the quote, the backslashes disappear anyway.

To quote the actual source with its format, links etc. intact you have to use “Quote whole post” when replying.

e.g.:

@LowLevel

In this example I wrote exactly \* and it will be shown just as \*

You can also insert preformatted text inside any normal sentence, like this

When quoting selected text all formatting is lost.

[PiecesOfKate]

@Guga

You have to escape your asterisks

Write * and you’ll be fine.

Ah yes. I tried doing that but had my backslashes facing the wrong way

@Guga

It doesn’t seem to make a difference in the quote, the backslashes disappear anyway.

That might be because it only turns them into bullets at the beginning of a line.

The problem with preformatted text is it strips everything including the font, which looks odd.

~Maybe I’ll start using tildes instead~

[Guga]

@Nor_Treblig

you have to use “Quote whole post” when replying

Where is it? I can’t find this option anywhere.

[ZakPhoenixMcKracken]

@Guga

note that I had to escape that backslash too, so to write * I actually wrote \, and to write this \ I had to write \\*, and to wr… well, I think you all understood.

Eheh, Viva la programmación!

[ZakPhoenixMcKracken]

@LowLevel

Do you think that posting in this section a short cheatsheet of the basic MarkDown syntax would be helpful to others?

I think so, please do it.

[milanfahrnholz]

@Guga

Where is it? I can’t find this option anywhere.

It´s the one that says “quote whole post”

Seriously though, it´s the speech bubble left of the “Bold” icon.