I've been wondering lately about whether comparing link attributes in HTML code would provide any value to the search engines. Testing's been done by SEOs in the past about the value of link titles (minimal if any) and rel attributes other than nofollow (nada), but I'm wondering about the target attribute, specifically. Currently, according to W3C, you can assign:
- _blank - the target URL will open in a new window
- _parent - the target URL will open in the same frame as it was clicked
- _self - the target URL will open in the parent frameset
- _top - the target URL will open in the full body of the window
Since parent, self & top are all, largely on-site options, I'm thinking that "blank" is the only option that could have any real impact.
If I were a search engineer, I'd definitely test out whether assigning less value to (external) links that open in a new window produced better results. My guess is that it doesn't mean a thing, but I know there's some smart people posting in the comments here, so I'd love to hear what you've got to say.
p.s. Note on the W3C schools page, the "nofollow" attribute doesn't exist in the "rel=" options. :)
I realize I'm slightly hijacking the topic - but why not then consider a class attribute for < a > if its name is something like "external"? there's been a lot of talk of Google paying attention to funny class names (e.g. "hidden" etc.). But like G-man says, as soon as something is implemented there are ways to abuse it.
I was a newbie to Google Adsense not long ago and I emailed them asking how to set the adsense links to "blank". I got an email back saying something like, we do not recognize this email address remind us who you are? ;)
It may be worth noting that the target attribute is invalid for strict DTDs. That's not mentioned in order to imply that search engines prefer valid code (I don't see any reason to believe they care as long as the spider can get through), but I doubt very much they'd put any weight on something that valid code can't include.
I highly doubt it would matter. There are so many easy methods to scam this approach that it would be next to useless as soon as it was implemented.
About six months ago I changed links on several dozen pages from no value to _blank. I had monitored these rankings for a long time at least once per week. No obvious change in rank was seen.
Rand, IMO you should change the links on your blog to open in new window.
I've used the value "_blank" internally within a site on a number of pages without seeing any apparent significance in difference from a pagerank stance when it was used, and when it wasn't - though it's difficult to tell from a ten point ranking system on a toolbar.
From a crawling perspective, if a spider had to check to see if there was a "target" attribute and what value it had defined, it would add to the time and processing power needed for the crawl.
So, there would probably need to be a meaningful reason to look at that value. I'm not sure that I see one, though. I don't think that you can assume significance to whether or not a page appears in a new window or not.
A quick search around Google Scholar doesn't show many papers that seem to discuss crawling and target attributes. The ones that do appear to use "target attribute" in a different sense than the html code.
I also looked through some papers from Junghoo Cho, whose writings (more here) on crawling have been pretty influential.