Linking collections objects to publications

1. Persistent identifiers (PIDs) that we know and love

DOIs (Digital Object Identifiers)

Academic articles have a DOI (e.g., 10.1371/journal.pone.0053873), as does the literature cited by that article (mostly).

  1. Reader can trace the provenance of ideas, data, quotes, etc.
  2. We get metrics of use (number of citations of that paper)
  3. Enable third party ecosystem (e.g. altmetric "donut")

2. Can we do this for collection objects?

Imagine we had a network linking collection objects (such as specimens in natural history collections) to publications on those objects. How could we build this? We need at least three things:

  1. Identifier for collection object (e.g., stable URL for herbarium specimen)
  2. Identifier for paper (or dataset, etc.) that mentions that object (e.g., DOI)
  3. A statement saying "this paper mentions this object"

3. How do we show these connections?

Let's assume that we have a list of specimens and the papers that mention them, how do we show those connections in a useful way?

  1. Try and convince academic publishers, museums, herbaria and other archives to add this to their web sites...
  2. ... or be like altmetric and just do it!

4. Bookmarklet demo

Annotate It! ← Drag this to your bookmarks bar then visit the following pages and click on the Annotate It! bookmarklet.



5. Summary

What's Bin Did

Using a bookmarklet we have displayed links between academic publications and objects in natural history collections. In each case neither the publication's publisher nor the natural history collection knew about the links, but we were able to add them to the web page using a bookmarklet. We can enhance the user's experience, and gain insight into connections between objects and research based on those objects without requiring changes in the underlying web pages.

What's Bin Hid

Under the hood there are two key components, the bookmarklet that reads the web page, and a triplestore that stores links between collection objects and publications as W3C annotations.

6. What's next?

Is this a useful way to show the potential for PIDs for collection objects?

If so, how do we scale this across collections that already have PIDs?

Can we generalise this beyond natural history collections?