Canonical Link Header element - Pointing from PDF to HTML?

[bakedjake]

From the manual [support.google.com], Google gives us a way to set canonical tags on PDF documents via HTTP headers, useful if you use multiple versions to track:

you can use a rel="canonical" HTTP header to specify to Google the canonical URL for the PDF file, as follows:

Link: <http://www.example.com/downloads/white-paper.pdf>; rel="canonical"
Google currently supports these link header elements for Web Search only.

Anyone ever tried this pointing at an HTML document instead?

So let's say I have a PDF of my HTML content. And the PDF is a mirror image of the valid HTML content, just in an easily viewable form for offline consumption. Both the HTML and PDF file are indexed and rank.

What happens if I serve the canonical link header on the PDF file pointing back to the HTML file? Will Google treat the HTML version as the canonical version?

[Walt Hartwell]

As I recall, the canonical page should be a very close, if not exact, version of the duplicated page. The example I remember was "the sort order of products may be different, but overall it should be similar". I would think a PDF to html canonical tag wouldn't fit in that kind of category. I haven't tried it, but I'd be interested in results of such an experiment.

[Walt Hartwell]

Just some correction as I read a bit about it today. PDF to html canonical is considered perfectly fine by Google.

[Nutterum]

Yes, PDF to HTML is absolutely fine and some websites actually adopt it as it provides better UX for the same rankings (better linking opportunities, or internal links to other pages, etc.)

[bakedjake]

Great, appreciate the input. Thanks!