rel=nofollow

posted on 2005-01-20 at 23:19:56 by Joel Ross

In case you've been hiding in a hole lately, you've probably seen something about the latest attempt to stop comment spam. Apparently, the major search engines (Google, Yahoo, and MSN) will support a rel=nofollow attribute in the href tag, and will not index those links. This means that spam by comments (like I get on .Text blogs all day long) won't get any Google juice.

Now, will this stop comment spam? No. Not in the short run. In the long run, when all blog software supports it, then it might. But until then, it will continue, and even though the spam won't be as profitable as before, there's still the chance that they will hit someone who hasn't made the change. And not everyone will switch. .Text 0.95 will be out there long after Community Server is released, and not everyone will implement Scott's change he posted.

So what is the best solution for comment spam? I thought at one time CAPTCHA was, but I recently read something that reminded me that, when I used RssBandit, I could post comments directly from there. Now, with CAPTCHA, no one can comment on my blog that way. They have to come out to my site. Which isn't necessarily bad, but for me, my blog isn't about driving traffic to my site; it's about reaching and communicating with the most people as possible, regardless of how they get the content or respond to it. Adding CAPTCHA hinders that process. On the other hand, since implementing it on Tourney Logic's blog, I've had no spam at all.

There are other ways are to prevent comment spam. One is to run an unknown blog software. Not much comment spam on my b2evo blog. But that's not a real solution. In .NET, you can dynamically create the controls needed for comments, using random ids. That prevents posts without those unique IDs from being posted. I'm sure other languages could do the same type of thing too.

Anyway, the new nofollow tag has also created an interesting side effect. You now have the ability to prevent others from benefiting from your Google Juice. Scoble brings up the example of linking to competitors.

You see, Google assumes that when you link to someone, they are a trusted source. That's not necessarily the case, and with nofollow, you can tell Google just that. For example, when Engadget linked to Kryptonite's site, it wasn't because they were providing top notch products. It was because you could pick the lock with a Bic pen. But how did Google know it was a bad link? Answer: it didn't. But now it could have. Is that good or bad? Is it an abuse? Will this help or hurt Google's Page Rank? And I don't mean my page rank, I mean the page rank system. Google is allowing the public to manipulate their internal ranking system. Will that hurt it's reputation?

I guess time will tell.

One other bad thing about it. It already prevents legitimate commenters from getting any benefit from crawling. So what will happen? Well, bloggers will start creating posts pointing to other posts rather than just commenting. Didn't Scoble recommend that once anyway? And people who comment for Google juice alone will stop (that's a good thing). Of course, this could be solved by allowing bloggers to flag comments as trusted - still show all comments, but allow you to flag a post as trusted, removing the nofollow attribute from the post.

Anyway, read around. There's lots of talk about this right now, as it's a hot topic. I'm still undecided on it. At first glance, I was all for it. The more I've read and understand the implications, the less I like it.

Categories: General