Geeklog Forums

Any thoughts on Google's new anti-comment-spam tag? Is it worth incorporating into Geeklog?

---

Arthur (http://www.shrednow.com)

I guess this is more a question if you are a blogger, famous blogger or even provider.

It surely sound interesting. However, the question what google thinks about the comments on my webpage is truly secondary to me. Much more important to me is not to have any spam on my page, if google follows the links or not. I dont want the users to see them in the first place. So there will be no other way than to actvate notification on comments and delete all the spam.

If, however, you have such a large site that all the comments given are so many that you do not want to read them all, and if you do not care about a small number of spam amongst a HUGE number of true comments, this might be a good idea.

If you are evne a provider and simply do not want to care about the thousands of blogs runnning on your servers, this might bea good way to ensure that search engines do not blacklist your servers in any way.

I guess it depends on the purpose of comment spam. If it really is to artificially inflate search engine rankings (the commonly accepted theory), then this could work - after all, who's going to waste their time comment spamming links when the links won't boost your rank?

If the intent is to get people to click the links, then no, it's not going to do much good.

But since I'm not a spammer, and I've yet to read one which is the real purpose, I say why not try it (assuming it's easy to work into the next release and doesn't compromise anything)?

The comment spammer's use automated bots to post the comments. They use search engines to find their prey.

This needs to be implemented in GL, It would not be hard

---

I work for a Wilmington NC Attorney

I have to say I'm not too impressed. What SixApart and others seem to deploy now are hacks and plugins to tag that new attribute onto all links in user-contributed posts. Obviously, this would also apply to "valid" links. So, in other words, in order to devalue the spammer's links, we devalue all the links? Can't say I like that idea ...

Having said that, I can see this attribute being useful, as it would now allow me to link to a site that I disagree with, say, from a commentary piece about that site. So I think the idea is okay, I just don't agree with the way it's being implemented.

I also don't think the spammers will care too much (at first, at least). There are enough sites out there that aren't maintained any more or can't or won't upgrade their software for whatever reason. The spammers have enough targets and enough resources at their disposal.

For example, geeklog.net is being hit by spam from one particular source several times a day. Yet, 99% of those posts are caught by the SpamX plugin and don't even make it onto the site. And the remaining posts are removed manually within a short time. This has been going on for months now, but that hasn't stopped that person. They probably haven't even noticed it yet.

I've read quite a few comments and blog entries that hail the rel="nofollow" attribute as the end of comment spam. I'd suggest reading Mark Pilgrim's rather grim view of the situation and the future of comment spam to put things into perspective.

bye, Dirk

The above post was my personal opinion on the matter. So, from a more objective and technical point of view: What can be done to support this new attribute in Geeklog?

The first thing to do would be to add it to the list of allowed attributes for the HTML 'a' tag in your config.php:

You may also want to go through your theme's template files and apply the attribute here and there. Places that would come to mind are the link to the user's homepage in users/profile.thtml or maybe even the link details, links/linkdetails.thtml.

There are probably a few template files for which Geeklog only provides complete link tags so that you can't add the attribute yourself. If you come across such a case, let me know and I'll look into it.

The next Geeklog release is quite some time away, and I don't really think this is important enough to rush something out now. So anything that requires changes in the core code will have to wait.

Instead, I was about to suggest writing a module for the SpamX plugin to do this (i.e. simply slap a rel="nofollow" attribute onto all links) when I realized that the plugin API doesn't allow modifications of a post - you can only approve or reject it. Have to look into this ...

At the moment, I have no idea how this attribute could be applied selectively (other than manually adding or not adding it to links when editing a post), so I'm open for suggestions.

bye, Dirk

One way that I am continually getting spammed is through the referer links with the stats plugin. To add the tag to the stats block of the stats plugin where it shows referers do this: change line 620 of functions.inc from

to

Tom, is nofollow already implemented by Google even though it's not officially supported?

I mentioned on groklaw that I think it could realy help them there. PJ has said that one of the big reasons she blocks google and the other search engines is because of comment spam. If there was a way for her to deal with comment spam then lots of good information about the unforunate sco incident could be made searchable.

IMO, comment spam should be filtered out as much as possible - that's what the SpamX plugin is for.

Of course I realise that Groklaw has "slightly" more comments than we have over here. Is this a server load issue?

bye, Dirk

This is an important feature. It eliminates the motive for a lot of comment spam, stopping it at the source. Of course it will continue on for awhile. The nofollow thing will result in a decrease in spam over months, not tommorow.

As for Groklow, why not just deny the a tag in comments on that site? That's kind of extreme, but it seems a lot less extreme than blocking google! People can post links that can be cut and paste into a browser if they want, and of course story admins could make any HTML they want.