We integrated ROR into Janeway in 2024-25
The Research Organization Registry is “a global, community-led registry of open persistent identifiers for research and funding organizations.”
User stories:
“As an author I want to create and edit affiliations for me and my co-authors.”
“As an editor I want to edit affiliations for authors.”
“As an author or editor I want affiliations to be clear to readers and funders.”
Searching to add an affiliation on Janeway
ROR on Janeway article page
ROR in Janeway-generated Crossref metadata
Hold on. Why?
We now have some links:
Why #1
Why do we need them? Can’t authors just type their affiliations in and call it a day?
Five whys: ask why five times to discover the root cause of a problem.
Griff, my college English professor: “Why is it thus and not otherwise?”
The Web uses URLs
URLs are a ridiculously powerful vehicle. Their distinctive powers are:
- universality
- location
They can go get anything from any context (in theory) and put it next anything else.
Web inventor Tim Berners-Lee's catchphrase "THIS IS FOR EVERYONE" at the 2012 olympics.
Why #5
I asked you about names. Why are we talking about links?
Principles that shed light
In 1990 Jakob Nielsen began developing usability heuristics for interaction design.
Today "Nielsen's Ten" is still used as a checklist for noticing what's wrong with a user interface, and what's working well.
Users search with the name they know
“Well, I know it has Washtenaw in it.”
Users don't edit the ROR ID or display name
“Oh good, someone has already put it in.”
Users are shown other info they may recognize
“Ah yes, that's the WCC I know.”
Users can put ROR ID in Janeway using ORCID
“It looks like it got my affiliation info from my ORCID profile.”
URIs combine to create linked open data
We can get ROR IDs programmatically from ORCID because they encode them as URIs, creating linked open data.
- User registers with ORCID
- Janeway checks their public ORCID profile for affiliations
- Janeway imports the primary affiliation and matches any ROR URI with other ROR data in the Janeway database
- Janeway shows the post-login screen to the user
Users expect us to talk to other systems
“I have already entered my affiliation on ORCID and made it public. Can you just use that?”
Users may want to edit or delete the affiliation
“Ope. My affiliation on ORCID was out of date. Let me just change that.”
Turning to ORCID IDs
People’s names overlap and vary in form, just like organization names.
Real examples from orcid.org:
| First name | Last name |
|---|---|
| Joseph | Muller |
| Joseph | Muller |
| Joseph | Muller |
| Joseph | Muller |
| Joseph | Muller |
Should use a search interface, like with ROR records?
A search interface to choose an ORCID is not always enough. It is better to have each ORCID holder authorize the link by logging in to ORCID.
Users can log in with ORCID
“Yes please. One fewer password to track.”
Oh dear! ORCID IDs can be saved in various ways
Here are some real ORCID ID field values for published OLH articles:
| ORCID ID |
|---|
| 0000-0002-2911-8382 |
| https://orcid.org/0000-0002-9651-3315 |
| orcid.org/0000-0001-9039-2201 |
| https://orcid.org/ 0000-0001-8731-9097 |
| ORCID: 0000-0002-1148-689X |
| 000-0002-8489-4405 |
| 0009-0002-9526- 9012 |
| nasukawa |
Anything can be entered into the ORCID ID field, with no required format.
One result: Janeway might not know which ORCID ID format to check for, letting you create a duplicate account.
(Lest you get too worried: Janeway does makes ORCID IDs uniform in outputs.)
Users never enter ORCID ID manually
Work in progress by Esther Verreau and Alainna Wrigley at CDL:
This small button will make a big difference to users and improve the quality of ORCID metadata.
Turning to DOIs
Article titles vary and overlap too, so the five whys are similar for DOIs too.
Real examples from Crossref:
| Title |
|---|
| Back to the Future |
| Back to the Future |
| Back to the Future? |
| Back from the Future |
| Taking back the Future |
| Back to the Future II comment |
| The Future Is Back; Back to the Future!* |
Differences in circumstances:
- Generation and registration of the URI rather than retrieval and linking
- Journal editors and press managers, not authors
- Additional namespace in URI path (e.g. 10.1234)
- Multiple registation agencies
- Location matters just as much as identity since a DOI represents a digital resource
Users can check resolution before publishing
To resolve, a DOI must be:
- A valid URL
- Registered with doi.org via Crossref or Datacite
- Pointing back at the article page on Janeway
Users can see how the DOI resolves and catch errors
This brings in two more usability heuristics:
What if: Users can check DOI validity earlier
We could add checks based on the expected character sequence of a DOI (using regular expressions).
These checks would happen before any communication with Crossref or Datacite, when the user...
- edits the DOI prefix and pattern fields
- advances an article, triggering a deposit
- manually creates or edits a DOI
