| 12:24 am on Nov 27, 2013 (gmt 0)|
I do even on sites that are canonicalized via htaccess without much chance of on-site duplication, just because if there's a lazy or forgetful scraper out there who copies a page(s) and forgets to change it then their copy is pointing to the original as the source.
Basically I go with the theory of:
"Can't really hurt, but could help, so why not?"
| 7:45 am on Nov 27, 2013 (gmt 0)|
Canonicalization errors can occur on static pages as well as dynamic pages. The rel="canonical" link provides some extra insurance. As JDT suggests, why not use it?
| 7:59 am on Nov 27, 2013 (gmt 0)|
Just to confirm, on each page I should only be inserting <link rel='canonical' href='link to page'> and nothing else?
| 9:16 am on Nov 27, 2013 (gmt 0)|
Yes, just make sure it goes in the <head></head> section or it will be ignored for security reasons [if it wasn't ignored in the <body></body> section someone could come along and drop it in the comments on sites that allow them, etc] and it's best to use a fully-qualified URL, so
<link rel="canonical" href="http://www.example.com/the-path/to-the-file.ext">
Edit: Very minor edit
[edited by: JD_Toims at 9:29 am (utc) on Nov 27, 2013]
| 9:20 am on Nov 27, 2013 (gmt 0)|
Excellent, thanks everyone!
| 1:46 pm on Nov 27, 2013 (gmt 0)|
sorry - could I just ask one more question?
We <meta name="robots" content="noindex"> our 'cart' page as we have no need to index this and also want to prevent any https pages being indexed.
Although this is 'no-index' should I still use the canonical tag for the http version?
(hope this makes sense!)
| 7:02 am on Nov 28, 2013 (gmt 0)|
the link rel canonical is an indexing suggestion for google, so if you have specified a meta robots noindex element, the link rel canonical is irrelevant.