Google now supports link rel="canonical" in HTTP headers | Hacker News Reader

Google now supports link rel="canonical" in HTTP headers

googlewebmastercentral.blogspot.com

28 pointsnephics5 comments

The first comment has it spot-on for the first example: the html version is an alternate version -- it's not the same content. The example of content distribution networks (and other mirrors) is much better.

The syntax is very strange.

rel-canonical: http://www.example.com/white-paper.html

would have been a better one.

The Link: header has been a part of the HTTP spec since forever. The following example from the HTTP/1.1 spec [1] should look familiar:

    Link: <http://www.cern.ch/TheBook/chapter2>; rel="Previous"

We should be applauding Google for supporting standard headers instead of making up new ones. The Link: header is even mentioned in the HTTP 1.0 spec [2] and, as mentioned in the OP, there's even an RFC acting as a registry for the various Link relation types [3].

Though I don't always agree with standards, I hope you'd agree that it really does make sense in this case...

[1] http://tools.ietf.org/html/rfc2068#section-19.6.2.4

[2] http://tools.ietf.org/html/rfc1945#appendix-D.2.6

[3] http://tools.ietf.org/html/rfc5988#section-5

What's the difference with the <link rel="canonical"> tag ? For those documents that are not HTML ? Seem not very useful to me.

It works for both HTML documents and non-HTML files. For non-HTML files, you cannot use the HTML link rel="canonical" tag, so this new support for the header is the only option.

Reply on news.ycombinator.com