Google's new rel="canonical"

Hi!

I just put my new site online, but had some problems with duplicate content (I suppose).

Then I found this post from Google, explaining the new [tt]<link rel="canonical" href="...">[/tt] feature to explicitly reference the original page (adding SEO ranks to the original rather than the parsed page): http://googlewebmastercentral.blogspot.com/2009/02/specify-your-canonical.html

What would be the way to insert this into MODx? Should I patch the tag select menu from the META tags and keywords?

João

->Edit: explain canonical link

I think that this could be done by editing the StrictSEO plugin. BUt since that redirects to a single url it defies the purpose. Then maybe the virtual alias plugin can do this. The easy way would be to add a TV and manually assign the canonical reference.

After thinking a bit abuout this new feature, it seems that the search engines have simply given up on making it their responsability to identify duplicate content based on url string. How hard can it be to know that domain.com/index.php?id=1&color=red is the same page as domain.com/index.php?id=1&color=blue but with a parameter change? Just read the content and see if it is the same and choose one (preferably without the color= parameter). I guess it is a sign of the economic crisis and they are collectively redirecting their efforts. Still seems like a cop-out.

Something simple like this in the main template:


?

Quote from: mrhaw at Feb 16, 2009, 02:57 PM

Something simple like this in the main template:

<link rel="canonical" href="http://www.example.com/[*alias*]" /> 

That makes it look easy and is a good idea. It will work if the document has an alias and is not using the "Use friendly alias path" option. You could use a snippet to check if alias exists else use id just in case not every document had an alias.

Quote from: dev_cw at Feb 16, 2009, 02:45 PM

... Still seems like a cop-out.

A cop-out in the good sense, surely. It seems a good move by Google to start and support this kind of URL - after all, a lot of sites rely it

Quote from: garryn at Feb 16, 2009, 06:02 PM

Quote from: dev_cw at Feb 16, 2009, 02:45 PM

... Still seems like a cop-out.

A cop-out in the good sense, surely. It seems a good move by Google to start and support this kind of URL - after all, a lot of sites rely it

Don’t get me wrong, it is a good thing. I was just saying that the Search Engines gave up on resolving duplicate content on their own and shifted the work to the sites themselves, which is a good idea (although the extra markup is arguable).

Is there still time to include a canonical placeholder in Revolutiion (make it standard)?

@dev_cw: Yes or [~[*id*]~] can return alias by mod_rewrite if I get this thing right:
http://www.webmasterworld.com/forum30/28332.htm

I’m working with friendly URLs so using the alias won’t work.

From the suggestion of mrhaw, I created a text TV (named originalPage) with default value [tt][~[*id*]~][/tt] and added [tt]<link rel="canonical" href="[*originalPage*]"/>[/tt] to the template.

So whenever the page is to be canonicalized, it’s sufficient to write the link to the original one in the TV value form (eg, [tt][~34~][/tt]). And for original pages, nothing needs to be done.

This approach is friendly-urls-friendly, as the generated link is then parsed by MODx.

Thanks!
João

I was referring to a built in tv: http://modxcms.com/document-tvs.html
But your approach works for sure!

OK, I just had a go at the snippet mentioned by dev_cw, but it is not working for me.
Probably completely misunderstand how to do this, but if someone could point out how to do it, I’d appreciate it.

Created snippet canonical
<?php
$alias = $modx->documentObject[alias];
// if alias blank use the docid
if ($alias != "") {
$output = $modx->documentObject[alias];
} else {
$output = $modx->documentObject[id];
}
return $output
?>

Then I put

<link rel="canonical" href="[(site_url)][[canonical]].html" />

in the template