Forum Moderators: open
Not just Google, but all the major players in the search and blogging spheres. However, this is a Google initiative.
The specific method described is correct, and the rel attribute is designed to mark "untrusted links".
blog announcements:
[blogs.msdn.com...]
[sixapart.com...]
[ysearchblog.com...]
In other words, if context is required or highly useful, the best indicators of context are not the invisibles on the page like rel attributes (or meta tags), but the human text on the page and the HTML markup that renders page layout since that is what the SERPs will reflect, and is the "product" the engine's customer is looking for.
I do agree that this context will need to be used: I see nofollow as subtratcing from efforts to make search engines more contextually useful.
Not to mention, the search engines can certainly evaluate the usage of nofollow on a site and even make that a ranking factor. If things become too bad, they can even decide that they will ignore that nofollow and go ahead and use those links in PR calculations. But I doubt that it will come to that because its usage will be limited.
He, heh.
OK I'm stupid.
I'm stupid 'cause I know that betting against BigDave is dumb. But BigDave, I got $5 that says within a year sites will be getting filtered if not penalized for using this for SEO reasons.
I can't think of an easier way to spot an SEO'd site that to look for these tags in combination with a few other common SEO footprints. Doh! This could be the cleverest SEO trap to come along in a while. Not that it was intended to be that, originally. ;-)
No it won't.
I what you suggested is not exactly what I was saying.
I was suggesting that I do not think that it will come to ignoring any links that have nofollow as an attribute, and that google is unlikey to even consider it a factor in the general ranking algo.
What you are suggesting is using it as a specific way to spot those that abuse the system. That is a completely different creature than the general algorithm.
I would not bet against them using it as an easy way to raise a flag on a site. I would bet against them making the heavy use of this particular attribute a negative factor on its own.
Personally I do believe that sensible outbound links serve to improve a page's ranking. If that is true, excessive use of nofollow could hurt your own site's ranking without them having to change the current algo at all other than not counting those nofollow links.
I'm not about to use this attribute to hide any links that I am not already hiding from the search engines. It just makes it easier for me to do what I have already been doing.
Now, what others decide to do with it is up to them. But, if we look back at history, anything that is open for abuse will be abused. If you plan on using this to influence the PageRank of your own site by preventing the transfer of a vote from Site A to Site B, then expect some risks. What risks are involved? Who knows at this point. But, it sure sounds like there are a few of you who are going to find out, huh? ;)
This is definitely going to make for an interesting next few months while all the link development programs come up with a strategy to deal with the rel="nofollow" attribute.
I would not bet against them using it as an easy way to raise a flag on a site.
So they implelemnt the attribute in order to help out all the complaining bloggers, and then down the road, penalize the bloggers for over use of the attribute.
The whole thing is such a joke. It will do nothing in the imediate future to eliminate comment spam, because comment spam is about abandoned blogs. The active blogs get pounded just because they are in the way. No one is going to stop pushing the button because their success ratio has dropped a bit.
In order to fix it down the road, you would need all new blogging software to use the attribute by default, so all the future abandoned blogs become usleless. But doing that will ultimately destroy what makes the blogging community so powerful. And blog owners will have to spend even more time filtering out the crap, because they will now need to spend time building all their buddy lists.
And does anyone really think that the engines (especially Google) will actually be able to separate PageRank from anchor text using this attribute? They certainly haven't showed they have that ability up until now.
All in all, it sounds like a great idea, but I think it will ultimately create far more problems than it fixes.
BigDave, ah well, then we got no bet as I agree with most of what you just said. Darn, I was looking forward to making $5 a year from now. :-)
For me still comes down to this: the new tag just won't stop blogspam in any meaningful way, and it will be open to abuse of the kind practiced mainly if not exclusively by the more sophisiticated players. Not good for the little guy.
Still makes no sense to me...and has the potential to hurt the Web more than it helps it, for the innocents especially. It's gotten to the point that you can no longer find the cool little interesting content sites with a major SE. It is the way of the world, however.
The irony here IMO is that this will also reduce the presence (on the SE's) of some of the cooler sites on the Web right now: blogs. :(
<Added>WebGuerrilla, you beat me to it.</Added>
So they implelemnt the attribute in order to help out all the complaining bloggers, and then down the road, penalize the bloggers for over use of the attribute.
Did you even read what you quoted?
I was saying that it might raise a flag, not that it would draw a penalty.
You also might want to read how the blogging sites are planning on implementing it. The livejournal method makes a lot of sense.
Your post on your blog contains followable links. Your internal links, and links to topics and friens are followable. Links posted to comments by your own friends are followable. It will only be links posted in comments by those that are not listed as friends that will get the nofollow.
[webmasterworld.com...] msg #9
[edited by: encyclo at 12:46 am (utc) on Jan. 20, 2005]
Occasionally I'll find a site, or one will be suggested, that I'd like to link to because of the content, but one where it's also obvious that the business has used, or been sold, a package that includes tactics that are questionable.
If this could reduce my risk, that would be great.
But would that be an abuse?
We'll parse Javascript links and use them in our ranking algorithms. ;)
The surprise is that the engines came to together to work on this, (scary surprise too for some I would bet), and that comment spam, of all things, brought them together. Silence the most vocal critics first.
Blog spammers will probably still spam away in hopes that a user will actually follow one of their links.
Just my opinion.
Hey, raise your hand if you actually read the linked material (in the first post of this thread) before you flew off the handle... There's a lot of information and answers to questions posed here to be found.
For example, from Six Apart...
So, this solves comment spam?In a word, no. But it makes things a lot better. For both comments and TrackBacks, we believe this removes much of the incentive (in PageRank or relevance ranking) for creating comment spam. [...]
Where else could I use this?
[...] a list of referrers that’s automatically generated would be a great place to add the rel="nofollow" attribute.
How does this affect Movable Type template tags?
[...] In the event that a hyperlink already has a “rel” attribute, the “nofollow” relation is added to the front of the list.
...and from Paul, who is "an engineer working on search at Google," via the MSN blog:
Any link with this tag will indicate to a crawler it is not necessarily approved by this page and shouldn’t be followed nor contribute weight for ranking.
I'd like to paraphrase a few quotes and add my uppity responses...
> What if every webmaster implemented this?
Hey, what if every webmaster cobbled together at least HTML 2.0 valid markup? What if Paris Hilton became Pope? What if we stopped worrying about things that will never come to pass?
> Oh no, now I'll have to view the page source of link partners! That's a fairly complex additional step!
Hey, howzabout you start using recip-link checking software that's being actively developed? You know, the one that in its next point upgrade will check for "nofollow" attributes?
> The "little guy" doesn't have the knowledge, time or patience to deal with this.
Hey, knowledge is power. That's why you have a 50,000 page, dynamically-driven site that pulls in $20K a month and the little guy still uses the <BLINK> tag. The meek inherit the earth, the knowledgable inherit the web. And if you're so worried about the little guy, what do you do to help out your future competitor right now?
> This isn't going to wipe out blog spam, so why bother?
Hey, I bet you don't use anti-virus or anti-spyware software or spam filters, since they're not 100% effective either - so why bother?
Oh, how I could go on, but I have a life and a busted detector to fix...
Most blogs:
Put above points together, and you can find most of the comment links.
As if SEs cannot figure out comment links algorithmically.
Just suppose that they might consider ignoring all blog commentw to be inferior to ignoring only those that the site owner considers to be untrusted. Some blogs, and yes, even some forums, have good information in user submitted comments.
As it happens, some of the blogs that I follow have some of the best information I have found on certain subjects available on the web. And the moderated comments are generally even more useful than the articles. I think the search engines would rather let responsible blog owners decide on which posters are trusted and which are not
While being able to spot blog comment pages and guestbooks might be trivial, writing them all off is a less than ideal solution (not that this is ideal either)
In order to fix it down the road, you would need all new blogging software to use the attribute by default, so all the future abandoned blogs become usleless.
That's the whole plan, "to fix it down the road" and for "all new blogging software to use the attribute by default" for comments/trackbacks/referrers from unknown sources. Not for every single link.
If the plan was for blogs to use it on every link, then they wouldn't need link by link granularity, it would be accomplished with a HEAD robots metatag. If that had been the plan it never would have been adopted. The feature had to be built as a "per-link, default for untrusted sources" for it to be adopted.
Why wouldn't google just ignore the attribute on internal links as a matter of course?
I don't think Google or anyone else could filter links manually. There are too many variables. I have a site with comments, but it's all hand-coded. (I don't use MT or Wordpress etc.) I do have a header "Comments". But what if I took it out?
Also, you'd have to allow for thousands of possible languages that might be used to show the word "Comments". But the user might use a different word - or none at all - or an image. Etc, etc, etc.
I'd like to see all these groups that have come together forming an official body, say the Anti-Spam Group or something. Because they've agreed on something once so why not again in the future? We could see several ideas to halt spam. So long as the body includes all the major blog systems, we can assume they will implement each new idea with every fresh release.
If only browser makers were so friendly. We might have better support for HTML and CSS by now.
As for abandoned blogs, well they will always suffer. I know of forums which are running versions of software that has had many security upgrades. So they are wide open to hacking. I was once hacked when I didn't realise a new upgrade was out on a forum I had once.
When any new code or method comes along, abandoned blogs and forums will be even further behind. I doubt the web will stay as it is now. It's up to the webmaster to update their sites.
Current blogs merely have to wait for a new version of the software which will handle 'nofollow'. Otherwise, I'm sure they can do a temporary hack. They tend to run PHP or Perl which often gets modified.
I added the 'nofollow' attribute to my comments code last night. It took 2 minutes. I just had to type it in and copy and paste it. (Well, my comments don't yet allow for active hyperlinks, except for the user's name which links to their site. Any other links are just shown as text.)
My problem will be updating all the existing comments. I have converted the links in all closed comments to active ones manually. I will need to work on these over time. I doubt it will take too long. All part of site maintenance. I am forever tweaking the code here and there.
That's the whole plan, "to fix it down the road" and for "all new blogging software to use the attribute by default" for comments/trackbacks/referrers from unknown sources.
How does it FIX blog spam -- even down the road? Blog spammers DON'T CARE about the blogs where their links don't work, but they still hit them.
And bloggers aren't going to allow comment spam to stand just because the links aren't going to provide the linker with an SE benefit -- those comments (for active blogs) still have to be deleted.
Does the "nofollow" tag mean that the thousands of abandoned blogs are going to disappear? No. So there's still a target rich environment for spammers.
Maybe I'm particularly thick on this topic, but I don't see blog spammers being in the least bit worried about this Big Event.
Overuse of the attribute will *not* penalize a page. That would be silly as this will not stop linking on blogs (it wasn't meant to!), it will only stop (or at least help stop) spamming on blogs.
Some very important blogs in fact will continue to be very link heavy. The best ones sometimes are (hint hint Webmasterworld). The 'walled garden' approach has long been proven to be a bad idea.
After all, how is anyone going to tell? In fact, maybe even Google isn't going to fully implement this, maybe they're just trying to get people to describe their content better.
One thing to realise is that genetic algorithms may be in place. Google or whoever may not actually have a lot of specific policies on how they handle content, but rather let search engine feedback control the strength of different factors influencing ranking.
By having more factors at your disposal that you are measuring gives your genetic algorithms more of an opportunity to learn.
[edited by: blaze at 6:15 pm (utc) on Jan. 20, 2005]
Consider:
[speculation]
That G sees a certain ratio of "no follows" vs normal links as unnatural and this sets off a flag. This could then lead to discounting of the rest of the blog spam links that made it through without the no-follow attribute.
For example, if 65% of the links to your site are "no-followed" and on pages which are known to be blogs (as discussed above, this is fairly easy to detect most of the time) then it's a good bet G will penalise your site, possibly by discounting the value of your remianing (35%) links.
The idea is that if 65% of your links are determined to be spammy then it's a safe bet that the remaining 35% aren't totally legit either.
I don't know what the naturally occuring ratio of 'no-followed' vs regular links would be on the web as a whole,(65:35 is used just as example) but Google sure will be able to figure this out.
[/speculation]
