Hacker News Reader

Comment by pessimizer | Hacker News Reader

There's absolutely nothing spammy about http://www.metafilter.com/67307/Gampoumlmbampoumlc, and if it has to be excluded from the internet because it contains two links to the website of a product along with discussion of that product between two dozen people - your search is broken, not the page.

I'm not sure when it became alright to exclude "inane babble" from the internet, but I'm not comfortable with an algorithmic measure of lack of interestingness or quality being judged as an attack. I also think that it's moving the goalposts from detecting destructive or deceptive content to defining what all content on the internet should look like.

View on HN

That page is commercial spam (or "veiled spam"). It serves no other purpose than to get a link to his own product site, no discussion, no nothing. Just: For the special price of 999 Euro you can get your own. How is that quality content?

It is also spammy in that it doesn't nofollow external links on comments. Those links should have been nofollowed like 6 years ago. That they still aren't, is telling about MetaFilter's knowledge on SEO. You know that Google sees your site linking to bad sites, and what do you do? Complain about it? Fire people due to an update 2 years ago? Or do you simply take action to fix it?

MetaFilter doesn't have to be (and isn't) excluded from the internet. It just doesn't have to rank well for terms like "Gomboc".

It became alright to derank "inane babble" the moment it increased the satisfaction of search engine users. Which is pretty much from the beginning of search engines. It is user metrics that count. Statistics don't lie.

It does not matter that you are uncomfortable with an algorithm. I think I don't get your "judged as an attack"-part. The alternative is not using a search engine that relies on algorithms.

I did a casual investigation of MetaFilter. It has not changed much over the years. The pages I did look at were all of the level of "inane babble". One line comments. Puns. Not adding much to an already thin topic. Perfectly alright to have a place on the internet to debate how to get your friend to take Viagra. But not much need to rank that for men's health topics, relationship advice or even Viagra.

http://ask.metafilter.com/57406/Bright-light-flashes-in-my-r...

Now would you want to find that page when you are searching for those symptoms? Or would you want credible medical advice?

If it is moving the goalposts, it is moving the goalposts to better quality content. Those visitors not going to MetaFilter are not gone, they are catered to by other sites. Every tag on MetaFilter has a specialist site now. Ranking MetaFilter so high before 2012 was a present. MetaFilter doesn't deserve to be deranked any further, but it doesn't deserve huge pre-2012 rankings either. In my experience, todays SERPS are much better than ever before, and the complaints are biased, reminiscing good old times or influenced by the fact that some paid to post on that site (You'll defend it, as you don't want to back a losing horse, and your experiences with the site are probably good).

>It serves no other purpose than to get a link to his own product site

You've simply made this up. The poster (gleuschk) may have some undetectable connection to the product (just like you or I may have one), but it's the only post that he/she ever made about it. Judging by a skim of their thousand comment history on the site, what gleuschk seems to have is an interest in mathematics.

http://www.metafilter.com/user/2666

You could have figured that out with a click, though. Does the fact that you didn't bother make your comments inane enough to be filtered off of the internet as a spamlink for metafilter? Can you detect that with an algorithm?

Thanks for the link, though. I remember seeing that shape on a spam segment of a documentary on PBS a number of years ago, and it was nice to be reminded of it.

edit: From your examples of bad pages on metafilter, I know that I don't want you vetting what is or isn't quality content on the internet, yet I still trust you more than some associative algorithm.

Just for reference: metafilter bans self-links (ie links to content that the poster is involved with in any way) outright in posts to the main page. It's extremely unlikely that the poster has anything to do with the product in this post whatsoever. The fact that they're a member of fifteen years standing with only 15 posts to their name makes it even more unlikely.

I wouldn't argue that that this post deserves to rank highly in any search engine, but to downgrade the whole site because of it? Madness: It's a link to an interesting mathematical shape, with a bit on perfectly on topic discussion underneath which contains a few other relevant links. Best of the web? Probably not. Spammy? Hardly.

Links to spam sites will be flagged & then squashed by the moderators in short order. Well moderated comments by an active user-base shouldn't be nofollow - they contain actual useful information that a search engine (like Google!) ought to find useful, if page-rank matters at all.

It's possible that metafilter might have a problem with links in old posts ending up pointing to domain squatter sites I suppose, but if this is a major problem then it ought to show up in metafilter's google webmaster tools.

>metafilter bans self-links

I shouldn't have said that the poster owned that site, that was a mistake, and it detracts from my point. I should have left it as an example of a thin-content page that doesn't really add anything to the web. Of which there are many at MetaFilter. That the view that MetaFilter pages are all quality and deserve to rank well is not realistic. That it isn't a mistake to derank a site when user metrics and a/b tests show that that increases user satisfaction.

>but to downgrade the whole site because of it?

I do not think this happened. I do not think I found the single page on MetaFilter that deserved a downgrade of the whole site. It is but a symptom of a deeper underlying problem with the site's quality. Best of the web? Nope. Clear-cut spam? Nope. Somewhere in between? Yes. The rankings reflect that.

>Links to spam sites will be flagged & then squashed by the moderators in short order.

Except they aren't. Google reports followed links to spammy sites. I found a link to an Indian pharmacy in a few seconds of searching. Moderators ought to work. Nofollow will work.

The problem is not the old links, though they are accompanied by old stale content, which IS a problem. Also these are todays links:

- "Everyone On Wall Street Is A Dick."

- Is Using Lotion a Black Thing?

- #basketball #trickshots

These link to other websites like vine.co or b3ta.com. They are followed by some comments that new visitors can not join in on, lest they pay. This is fluffy content that ranked well around 2006, it doesn't anymore. Why are these comments so special that they deserve a top 10? When 100s of other sites also put up a link and have a small discussion? Where is the unique quality content of MetaFilter? In the comments?

Ah, I think you're operating under the misapprehension that it was metafilter.com that was attracting the Google clicks & so you're reasonably looking at posts to metafilter.com & wondering how they could ever have ranking in Google searches in the first place.

As I understand things, it's the ask.metafilter.com Q+A sub-site that was the main source of Google search clickthroughs & thus ad revenue (Yup, 90% according to Matt Haughey's essay on medium.com: https://medium.com/technology-musings/941d15ec96f0 ). The main metafilter.com site is pretty much irrelevant in revenue terms, although it may be affecting the overall ranking of ask.metafilter.com pages in Google search results for internal Google voodoo reasons of course.

NB. Let me know where that link to the Indian pharmacy is & I'll flag it.

Reply on news.ycombinator.com