How to use rel canonical element (tag)?

Matt Cuts from Google made a great new video explaining how to properly use rel canonical. One of the most often questions was what is rel canonical, is it a meta tag? Matt Cuts explains that it’s easy to confuse it with meat tag as it’s included inside the website where meta tags are usually placed, however since it’s actually a link element it should be called an element, not a tag.

The video shows us a few examples of where a canonical element should be used. Canonical element is used to eliminate duplicated content inside a website. A best example would be having a print version of article indexed by search engines. Print version is actually a duplicated version of your real article and you might not wish search engines to spider it and people to land on such pages as they are not made for visitors. Here comes the canonical link element which can be placed in header of a website and point to a original copy of the URL. In this case, you would place rel canonical on your print page with link pointing to the original article. This way search engines will understand which copy of your content is made for users and should be indexed and which copies serve for other purposes.

Matt also answered some “advanced questions” about canonical element:

Q: Can canonical element be used cross-domain (as 301 redirects can)
A: No, you can not use canonical element to point to original article on another domain. This element is to be used to eliminate duplicated content inside a website.
UPDATE: Rel=canonical can now be used cross-domain.

Q: Can we then use it with subdomains?
A: Yes, you can use canonical element to point an copy of an article on a subdomain to an original article on another subdomain or some other URL inside the same domain.

Q: Can you use link rel canonical for redirecting the same content from HTTPS to HTTP?
A: Yes, if it’s the same content it’s a great use of canonical element.

Q: What is the difference between canonical element and 301 redirects?
A: There is really not much difference between these two, except canonical element is restricted to a single domain while 301 redirects work cross-domain. Also sometimes people are not technically able to create a 301 redirect and there’s a point where canonical element kicks in.

Q: DO the pages have to be identical?
A: No, they do not, but they should be pretty similar.

Q: Can search engines follow a chain of canonical redirects? (one page pointing to another and that page to another and so on until original article)
A: Yes, but don’t count on it. Best use is to point all the copies of the content to the original article instead of creating a chain.