Calendar
Network
Feeds
ianmacfarlane
Hi Danny,
Rather than rewrite from scratch, I think your efforts would be better to put into making contributions to the Pligg project - it's open source and most OSS projects will welcome contributions with open arms. Contributing can be getting one of your employees to join and contribute to the project, or an alternative is specifying some behaviour you want and offering to pay the existing developers to implement it. This has the added advantage of helping other people out with this problem, and will help reduce spam on other sites too.
Just a note - this particular attack you're describing (with the multiple new account signups) sounds like an automated spam attack to me rather than done by humans (though you never know).
One of many possible parts of the solution could be site blocking - if someone spams your forum too many times with links to a particular domain (and the articles are off-topic spam, so obviously tied to the commenting, to stop 'Sphinn-bowling'), block all submissions and comments with the URL in it. Here you'll want to make sure you put in a written and obvious warning when submitting as to the reason it was blocked (so people can't comment about how awful the spam about spamsite.com was before and get accidentally blocked, and so people can also file for reconsideration as mistakies are inevitable in any system).
Another vaguely related comment with regards to the nofollow situation - have you considered allowing your super-admins to be able to flag articles as dofollow even if they don't get enough Spinns? This could be 3 people to 30 people, depending on who you trust. Things don't necessarily have to be popular to be interesting or useful (currently it's somewhat skewed in favour of sites with big readerships like Search Engine Journal).
Final suggestion - a 'duplicate stories detector' like Digg has would be an extremely useful addition to Sphinn.
Good luck with the spam fighting!
Ian
Robots.txt is a different situation from meta robots - if the page is excluded via robots.txt, then this prevents search engines from crawling it in the first place. Search engines cannot give any values to links they have never seen.
It takes a bit of thinking, but you can see the argument being made (though it seems a bit backwards at first glance).
Of course, someone could take the same statistics and argue the opposite.
Benjamin Disraeli would just love the modern era.
Sphunn - Interesting take, although I disagree with the idea that Microsoft would bundle Google Chrome with Windows. You might also want to have a look at our take on Google Chrome.
What I'm seeing as the major barrier for Google to overcome is how to get this onto people's computers.
See also our take (Sphinn link).
What I'm seeing as the major barrier for Google to overcome is how to get this onto people's computers.
See also our take (Sphinn link).
Where he says to ignore the specs because they don't specify the correct format for meta keywords, he's incorrect (I've linked to my Sphinn comment rather than cross-posting it).
While I generally agree with this article, Danny didn't read the specs fully - they do clearly state how the meta keywords is written. In the specs URL he links to (B.4 Notes on helping search engines index your Web site) it says:
Some indexing engines look for META elements that define a comma-separated list of keywords/phrases, or that give a short description. Search engines may present these keywords as the result of a search.
It then gives some examples - although these are single keywords, the text above makes it clear that both words and phrases should be separated by commas (meaning that you don't put commas in the middle of phrases).
EDIT - Incidentally, what the spec says is the conclusion that Danny came to, it's just that where he says to ignore the specs because they don't specify the correct format that he's missed that they actually do.
« previous1 next »
Story: Crappy MP3 Sites, Comment Spamming & Enough Already