On the other hand what the hell is this (care of Greg Dickson)?

Odd BBC "sport site" link to carrentals.co.uk

This is a link to a car rental website in the "From other sport sites" section of this story today about the England game.

That link is one of the automated ones that are generated via Moreover for the BBC - but  it goes via a series of weird redirects to the site in question. The actual link is to: http://c.moreover.com/click/here.pl?z4032419548&z=1250248769 which goes via a 302 redirect (invisibly via http://www.bbc.co.uk/go/redirect.shtml?http://c.moreover.com/click/here.pl?z4032419548&z=1250248769?z4032419548&z=1250248769) to the carrentals.co.uk site.

There are two issues here:

• The BBC isn't linking out fairly. 302 redirects don't pass link value in Google's eyes. If the BBC links to sites, it should link to them properly not via weird links that don't count to Google. It promised to stop doing these odd redirected links in 2008. Why's it still going on?
• Also, why the hell is it linking to carrentals.co.uk as a sports site? The Moreover feed needs cleaning up.

Fortunately, the transparency of the BBC means we'll find out soon no doubt ...!

Adding this meta tag used to work only for URLs on the same domain, but this has changed as this diagram from Google shows:

You can now use rel=canonical cross domain

Google has announced today that it will support rel=canonical across domains - ie if you have the same content on more than one website, you can tell Google which is the main version you'd like it to index. It's particularly handy if you've changed domain names, and don't have a way to do server-side redirects on the old domain / server.

You can read more about the change to cross-domain rel=canonical here.

Maybe I can use it to get my original post on reporting restrictions for bloggers to stop being outranked by the duplicate version on the online journalism blog. *Scowls at Google.*

The problem with linking

Put simply, Google counts a link to a page / website as a vote for that page / website. So everyone who blogged about, and linked to, Jan Moir's article on Stephen Gately has, unfortunately, helped that particular article AND the Daily Mail website as a whole do better in Google's results. (According to Yahoo Site Explorer, there were 334 links to the original URL and 541 links to the URL after it was changed - that's 875 links in total)

Google will notice a lot more pages linking to the Mail, a lot more websites linking to the Mail - and will decide that the Mail and that Jan Moir article must be more important than they were the day before.

(You don't need to worry about linking in Twitter, though, as I explain below.)

You can see the effect here. A search for the letter 'a' in Google currently brings up Jan Moir's article first. (Update - it no longer does.) (It didn't used to. Fortunately, at the moment, that article is only on the second page of results for a search on 'Stephen Gately'. Notice, too, they've forgotten to update the meta description).

Jan Moir ranks 1st for a search on the letter 'a'

Here's how to get round it ...

The easy solution - don't link to it. Sometimes you have to, however - in the short term, you may need your readers to be able to read something to make sense of what you've written.

I saw someone suggesting a URL shortener like bit.ly. This won't work as, with most URL shorteners, Google looks at the ultimate link (read this post on the difference between URL shorteners' use of 301 and 302 redirects to understand why).

The best way to avoid your link counting as a vote is to add rel=nofollow to your links.

This means the links will still work as before for normal users - but nofollow tells search engines to ignore the links when calculating pages' importance - in other words, the link no longer counts as a vote. (Google has an explanation of nofollow.)

How to add nofollow

If you've got blogger or wordpress, adding this is easy. Once you've added a link as normal, just click on the HTML tag of the editor where you write your post - it looks like this picture in wordpress.

Add rel="nofollow" in HTML view

Then scroll down to where the link is. You'll see some text like this:

Jan Moir has written a <a href="http://www.dailymail.co.uk/femail/article-1220756/Why-natural-Stephen-Gatelys-death.html">disgusting piece</a> for the Mail on the death of Stephen Gately

The bit code in < > brackets is what makes the link. So you just need to add rel="nofollow" (with a space before and after) inside the bit of code that beings <a href="http: ...

So it now looks like this:

Jan Moir has written a <a rel="nofollow" href="http://www.dailymail.co.uk/femail/article-1220756/Why-natural-Stephen-Gatelys-death.html">disgusting piece</a> for the Mail on the death of Stephen Gately

Then click the other tab at the top to return to the normal view. You won't notice any difference there - but you can be happy that, in Google's eyes, you're not promoting whatever you're linking to. Result!

Don't worry about Twitter

You don't have to worry about linking to something in Twitter. Twitter automatically adds nofollow to all links that you tweet. (Thanks to Minifig for the point.) And, likewise, comments in blogs usually have nofollow added to links automatically.

Message about the original page on the new URL

I wondered what would happen to the Sidewikis written on the original URL, such as this one.

The original headline and URL were:

• Why there was nothing natural about Stephen Gately's death
• /femail/article-1220756/Why-natural-Stephen-Gatelys-death.html

The new headline and URL became:

• A strange, lonely and troubling death...
• /femail/article-1220756/A-strange-lonely-troubling-death--.html

The answer is that Google keeps the Sidewiki, and adds a message at the bottom saying "Originally about: Why there was nothing 'natural' about Stephen Gately's death | Mail Online"

So you can't defeat Sidewiki by just redirecting an old URL to a new one.

What this means is that yourdomain.com/page is NOT THE SAME PAGE as www.yourdomain.com/page. The URL without the www at the front is treated as a different URL by search engines, for instance. And if someone leaves off the www (on a link or when they type a URL), they may not get any page at all unless you've set your server up right (explained at the end).

Does your website work on the non-www version, or do you get a message like this?

9 sites that don't seem to know this

This ought to be fairly well known in this day and age. But click the following sites and you'll go to the non-www version of their homepage - and in each case, you won't get anything.

• Body shop: http://thebodyshop.co.uk
• W H Smith: http://whsmith.co.uk/
• Woolworths: http://woolworths.co.uk/
• Jessops: http://jessops.com
• Her Majesty's Revenue & Customs: http://hmrc.gov.uk/
• UK Trade & Investment: http://uktradeinvest.gov.uk/
• Companies House: http://companieshouse.gov.uk/
• Oxford University: http://ox.ac.uk/
• Open University: http://open.ac.uk/

Solutions to the problem

You can serve up your whole website on both the www and non-ww version of your URL, but this can give rise to duplicate-content problems. It's better to redirect the non-www to the www version (or vice versa).

• To help Google: Set a preferred domain in webmaster tools if your site appears on both www and non-www.
• To help users: Write your htaccess file to redirect the non-www version to the www one.
• Help yourself: Read up about canonicalisation and make sure that SOMETHING appears on the non-www version.

But don't just leave awful messages or blank pages ... (and in case you're wondering why it's 9 sites, it was 10 but http://johnlewis.com seems to be working now!)