After my book, Bit by Bit: Social Research in the Digital Age, went through Open Review, I was excited that the First edition was finally ready to be printed and posted online. Then I ran into a problem. How should I post the First edition online while preserving the Open Review edition and all of the accompanying annotations? In this post, I’ll explain the problem in detail and describe what we ended up doing.
If you want to skip all the details, here’s what I recommend for other authors using the Open Review Toolkit:
- Build versioning into your standard url structure (e.g., https://www.bitbybitbook.com/en/open-review/preface/ and https://www.bitbybitbook.com/en/1st-ed/preface/).
- After you post a new edition, use a sitemap to tell search engines which version you want people to visit.
- After you post a new edition, add a note to the top of all pages from previous editions pointing readers to the most recent edition.
If you are interested in the gory details, keep reading. . . .
When we first posted the Open Review edition of Bit by Bit we did so without building any versioning into the url structure (e.g., https://www.bitbybitbook.com/en/preface/). Then when it came time to post the First edition, it was not clear where we should put it. We could have posted the First edition at https://www.bitbybitbook.com/en/1st-ed/preface/, but we were worried that we might lose all the “search engine juice” that comes from links going into the Open Review edition (e.g., https://www.bitbybitbook.com/en/preface/). After all, building up links to improve search performance is one way that Open Review can lead to higher sales.
The standard way to handle this problem is to canonicalize things using the rel=”canonical” link tag (see also https://support.google.com/webmasters/answer/139066?hl=en). Basically, by adding this tag, we would be telling Google and other search engines that the First edition is the “real” edition and that all links going into both the Open Review edition and the First edition should work together to point to the First edition.
So far so good, but this creates a problem with hypothes.is. If we use the rel=”canonical” link tag, then hypothes.is will move all annotations from the Open Review edition to the First edition. This makes sense for most use cases with hypothes.is, but that’s not what we want here; we’d like Open Review annotation to stay with that edition and First edition annotation to stay with that edition.
After lots of exploration and help from hypothes.is, we’ve decided that there is no perfect solution so we came up with the following solution for Bit by Bit (which is a model for the more general solution suggested at the top of this post):
- Posted the First edition at http://www.bitbybitbook.com/en/1st-ed/preface/
- Added a header to every page in the Open Review edition telling people that they should read the First edition
- Told google the canonical url through a sitemap (https://support.google.com/webmasters/answer/139066)
Pros of this approach:
- preserves Open Review edition with annotations
- allows separate annotations on the First edition
- easy to do
- easy to maintain
- easy to generalize when we have a second edition, third edition, etc
- might pay search engine penalty because of apparent duplicate content and/or not having our search engine juice flow from the internal pages of the Open Review edition to the First edition (or across future editions). However, we don’t have a huge number of links to the internal sections of the book (most are to the homepage) and we are going to mitigate this potential damage with a sitemap.
Now that this solution has been in place for a few months, it seems to be working pretty well. But, if you want to avoid thinking about all of this stuff, I’d recommend just following the advice at the top of the post.
I’m grateful to Luke Baker from Agathon Group and Nate Angell from hypothes.is for their help thinking through this specific situation and developing a better solution going forward.