Hacker News Reader

Comment by vadfa | Hacker News Reader

What I find more troubling is that amazon allows unmoderated user input to appear as suggestions.

I've had a bunch of bots doing oddly specific nonsense searches in my search engine, stuff like casinos, online pharmacies, warez. The bermuda triangle of dodgy internet sites. They are very strange because they are both generic and specific at the same time.

Here is an example:

> skyrim english voice files (and subtitles) latest version

You get no results with my search engine, but interestingly, if you feed that into google, you get some pretty dodgy top results.

It looks like the type of SEO spam nonsense you find on websites, hacked wordpress deployments are often rife with these types of keywords. I think they are trying to manipulate a would-be suggestion algorithm somehow. Problem is I don't do anything with the user input.

They were probably trying to post spam and thought your search bar was a comment input field.

Could be, but usually bots that post spam include some form of link, not just keywords. The link is the entire point of posting spam.

There also seems to be a few distinct bots doing this, with distinct user agents (but bots with the same UA spam-search on the same topic out of epharma, casino, warez).

I also have another, identical search box in another part of the site and it has never seen any sort of spam bots.

I wonder if these are attempts to drive up the “interest” metric, with the hypothesis that the more search for a specific string, the result that is exact match gets higher ranking?

Could definitely be something along those lines. Discovering all this black hat SEO stuff and what they are doing to manipulate search engines has been one of the great joys of this project.

Nobody seems to write much about it, maybe because they don't want to give people ideas, but it's an endlessly fascinating subject.

Seems fine as long as no one tells 4chan.

Whoops.

But also, interesting.

Reply on news.ycombinator.com