Hacker News new | past | comments | ask | show | jobs | submit login
Danny Sullivan: How Link Spammers Killed My Wife’s Web Site (daggle.com)
13 points by yarapavan on Oct 19, 2009 | hide | past | favorite | 23 comments



Sorry to be blunt, but stop bitching and fix the problems. I don't think this type of give-up attitude goes over well on HN.


What's even worse is I've used Drupal just enough to know that stopping link spamming is as simple as installing the right plugin. I haven't done it in years, but I'd bet it's even better now since there's almost certainly an Akismet plugin.

*Edit: just typed "drupal ak" in Google which auto-completed and took me to just such a plugin. I'm sorry, but if you can't figure that out, you have no business on the web.


A postscript by the author appeared:

"I’m not naive. I understand that sites should have anti-spamming filters in place. Lorna’s had some. It could have more. But I’ve also seen spam get through anti-spam filters on my own Sphinn social news site. That site employs multiple-CAPTCHA barriers, along with an array of other deterrents. It also has human moderators. Link spammers still attack it. Link spammers will attack ANYTHING out there, and nothing is foolproof. The first programmer that tells you they have a perfect anti-spam solution will soon after encounter another programmer who will blow that fallacy out of the water."

His experience pretty much mirrors my experience. Just using any of the main CMSs out there (drupal, joomla, wordpress, etc) will give you an entry in a bunch of databases of people that will 'visit' you with gifts from time to time.


In fact, Drupal founder Dries Buytaert has a side-project called Mollom for eliminating comment spam. And of course it integrates very nicely with Drupal.

Sounds to me like whoever they hired to do the site didn't have much of a clue.


They quite clearly said they couldn't be bothered to spend money on it as it wasn't earning anything.

She could, if she wants, get this latest huge barrage of spam cleared out on Boudica and improve spam filters going forward, if she choses. She’s currently debating this. It just may be, as I explained above, that this was the thing that tips her toward closing the site, something she was already considering because of low usage.

In a way it's probably not a bad thing if low traffic me-too style sites are discouraged a little - if there's something worth keeping then for the cost of a meal for two you can hire someone to install a captcha (or whatever) and strip out the spam. If the content isn't worth that then, does the web need it.

The story sounds more like the tale of someone who didn't have time or resources to do anything with there not-especially-unique website and then got annoyed because it attracted more bad attention than good.

The link-spam is an excuse, IMO, to close the site and not lose face. Sure it's annoying, sure it shouldn't happen (like people putting flyers on your windscreen), but don't lose sleep over it.


I just wanted to note that Mollom is free for small sites and takes no more than ten minutes to install (maybe two minutes if you know what you're doing) so yeah, I agree, it's more of an excuse for closing a small site.


I've had good success with the Captcha Riddler plugin for Drupal. It's just a very simple, non-obfuscated text captcha. It adds "type the word banana" (or any other question/challenge you want) to your registration form. Spam bots bounce right off it.

http://drupalmodules.com/module/captcha-riddler


Except it uses the Captcha module, which disabled page caching on any page that has a captcha -- a dealbreaker for high-traffic sites.

I'm using a hacked version of MyCaptcha specifically to avoid that problem.


Good point. But: there are a lot of spammers out there and their combined might can make anybody feel sorry about starting their website.

Some people walk in to this web thing totally unprepared for the assault by the bad guys and that can be psychologically enough to push someone to the point of saying forget about the whole thing.

Even with all the countermeasures in place I still find myself being very selective with the sites where I enable comments by anonymous users, even the best solutions are not 100% effective and when you get a few thousand or more attempts to spam shite some of it will get through. So it will always need a manual step to avoid things getting out of hand.

And in spite of all that, it's not rare to wake up to 1,000 or more comments using some new loophole that get posted in one night.

If his wife isn't capable enough to remove these from the database programatically then I can see how a few of those episodes would be enough to throw in the towel.

Spammers are seriously destructive.

Take a look at the 'new' page of HN with showdead on to see how much of it gets posted here. And quite a few of those are removed manually.


I see his gripe with it. There's nothing more irritating then seeing a whole bunch of spam comments on one of your sites. The worst is the compliment spam that is so hard to filter.

But what it's taught me is to be a better developer, and to learn new anti spam techniques. I feel bad because the site probably meant a lot to her. He should have helped her out more.


> He should have helped her out more.

I fully agree with that. I would do that for strangers, not doing it for your wife seems rather rude. It would take less time to put a 90% solution in place than it would take to write that blog post.


>He should have helped her out more.

It's possible she wanted to do it for herself without his help. FTA:

>She assembled the site entirely on her own, finding a programmer, working to develop the features and watching over the small community that made use of it.


Who defines good content?

Obviously blatant link stuffing on non-related topics is spam.

Obviously on-topic paragraphs of text on-topic without links is not spam.

What about on-topic posts with a link back to my affiliate sales on on-topic products?

What about on-topic posts with 5 links back to my affiliate sales page? (check out my favorite 5 workout videos on a post about womens health).

How substantial does the content need to be? "I really enjoyed this post, and I think that you are right on about XYZ. I read this book <amazon affiliate url>".

Basically, the problem with web spam is that the obvious stuff slaps you in your face as obvious. The gray area is really wide in some cases.

Also, the whole post was predicated on the fact that people have an obligation to provide you value when you put up a web form. That seems pretty presumptuous. I decided to send an HTTP POST with my content. You decided to post it up to the world.


That's a good point, but the vast majority of spam on my sites are clearly junk. If there were some easy way to stop all the obvious spam but left all the grey-area stuff, I would be overjoyed.

How about this: any comment posted by an automated process and not a human is spam?


If it's a human cut-and-pasting? Or someone employed to workaround strong captchas? Spam for sure.


Sure, and that's a growing problem. But 500 identical or nearly identical ads for fake viagara? I'd bet good money there's nobody pasting that into boxes over and over again.


As Justice Potter Stewart famously put it, "I know it when I see it."


"If Boudica had been more successful, doing the work would make sense. But it has remained small, and the link spam attack will probably tip her over to a decision she’s already been debating, of whether it makes sense to continue working at it."

In other words, link spammers didn't kill the site. They just pushed a decision she was already contemplating.


I just wish the spam bots were smart enough to identify sites that add nofollow to links. If that worked, it would be such an easy fix.


That would be true, except (in theory) the purpose of spamming isn't to put links out on websites, its to put links out on at least tangentially related sites, so that you can game the search algorithms better.

That means it doesn't matter if its no follow or not, you're still getting benefit. You may not be getting a direct benefit to your own site, but at the least you're hurting a "competitor".


I don't quite follow -- I thought a nofollow link doesn't enter into the search algorithm at all?


It doesn't (on search engines that "matter"), but if I sell say titanium hammers, there's still value to me putting link spam to my site on a construction blog.


Well, there are thousands of people out there whose sites aren't being killed by link spammers, so I think he should scratch his head and look home.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: