Come together with the global Drupal community in Rotterdam, 28 Sept – 1 Oct 2026. Sessions, contribution, connection, and Early Bird savings until 8 June.
By daniel read on
As I'm sure other Drupal webmasters can attest, despite best efforts, some comment spam still slips through and needs to be deleted manually. I've yet to figure out if this is from bots that have figured out how to break the captcha, or from humans who are adding the spam manually--but that's for another discussion.
What I wanted to post a wish for is the ability as webmaster to control more granularly whether or not the links within a comment use rel="nofollow". When a comment is new, it would be nice to have it use rel="nofollow", but then if I come along later and mark the comment somehow as "links approved", the rel="nofollow" would go away for that comment. I would prefer that this feature *not* be coupled to comment moderation.
The reason for requesting this particular design is that what happens is that some comment spam will slip in overnight while I'm sleeping. I delete it in the morning, but in the intervening hours, search engine bots have come along and indexed that page and added the spam keywords to the search index. So for several days after I have deleted the comment, the search engines are still sending porn and gambling searches to that page.
Another nice feature that would help in this area is a way to configure Drupal to reject a comment automatically if it contains more than a certain number of URLs pointing to a domain other than the "base URL." Most of the comment spam I get these days is packed with dozens of links, and a normal post would not contain that many links. It would be nice if Drupal were smart enough to help me with this. At least it would force the spammers to keep their comments shorter.
Thanks,
Dan
Comments
rel="nofollow" won't help
Dan,
rel="nofollow" won't help you to fight against spammers. They'll spam anyway and don't care about the rel attrubite. The second idea -- limiting the number of links in comments -- looks great.
Marek
Anchor Text Weighting and nofollow
Hello Marek,
Thank you for the feedback. However, I don't agree that nofollow doesn't help. My thinking is that googlebot, for instance, rewards anchor text; just as they don't confer pagerank if nofollow is present, it makes sense that they would give less weight to the anchor text in a nofollow link. If the nofollow is *not* there, and the search spider follows the link and sees that the remote site's keywords match the anchor text keywords, then it seems to that makes the keyword relevancy to my page higher. If nofollow *is* there, then does the spider even follow the link? And if it does, would it bypass any part of its algorithm that would compare the keywords?
The goal I am trying to achieve is to lessen the damage that the spam does for the short time it sits on the site. Maybe my theory of the effect of nofollow on the weighting of anchor text is off base, in which case I would be more inclined to agree with your statement that nofollow won't help reduce this particular variety of damage from comment spam.
Thanks again,
Dan
on-page vs. off-page factors
Dan,
What you are speaking about are so called on-page ranking factors (e.g. anchor text from the point of view of the linking page), while rel="nofollow" should affect just the off-page factors of the pages they link to (e.g. PageRank of the linked page or anchor text from the point of view of the linked page).
However, the most important thing here is, that those spammy links are very short-term. Even if Google finds them once on your page, next time it visits you, they will be away.
I think a spam prevention would be a better investment. That's why I agree with your suggestion to limit the number of links in comments.
(I'm sorry to be rather off-topic, but as a SEO consultant I had to add my two pence ;-)
Cheers,
Marek
Off Page vs. On-Page
Marek,
I don't think you're off-topic at all. I think this issue goes directly to whether or not my idea for "configurable, flexible nofollow for comments" is worthwhile. If you are indeed correct that Google does not and in the future probably would not add weight to in-page anchor text based on confirmation of the matching relevancy of off-page keywords, then the need for my feature to use nofollow for new, "unapproved" links is reduced.
I know the effects of the damage on the search engines is temporary, but I hate giving those bastards anything, even if it's just in effect for a few days. And I hate it when some old page on my site shoots up to the top of the site popularity ranking because of irrelevant porn searches in the aftermath one of these incidents.
Dan
Spam Module Blocks Based on Link Count
I wanted to follow up for people finding this via search that the spam module does have a filter for detecting spam posts based on the number of links in the post. It's a configurable threshold, which is nice. I must admit that since I had gotten so good at keeping spam out manually, and since I feared legitimate comments being marked as spam, that I had never used the spam module. I started using it recently, and I've been very happy with it. It has prevented a number of spam comments from getting through.
A few still slip through, which was the root issue in my original post. I would still love to have the "nofollow"-until-approved feature described above, though.
Best,
Dan
I was too thinking of being
I was too thinking of being able to force remove nofollows from selected nodes, primarily to reward the quality of the submissions. While it is a good idea to nofollow comments, it is also a good idea to remove nofollows from links you think contribute to your website.
Anyone working on this kind of module?
a step further
I think the steps should be:
1. no link is shown -> no spam value
2. nofollow link is shown
3. normal link is shown
An easy handling in the backend would complete this steps.
Robert
http://www.finanzen-online.info
The most interesting that
The most interesting that this site uses rel="nofollow" in comments.
http://browsergame.ru - the link is with this attribute
but which module can do it?
I need only the modul anyone
I need only the modul anyone know the name?
Mandy
http://www.aktiendepot.ws