Citation generator additions and changes

[jpastva]

Revised on 4/25/19 following DIWG discussions:

-Replace page header "Citation Formats" with "Cite this version of the work"

-Reformat the citations generated in DigitalHub to follow the formats listed below. Stop pulling data from Location fields. Labels with double square brackets should be pulling data from the item's metadata.

MLA:
[[Last name, First name]]. "[[Title]]." [[Date Created]], [[Publisher]], [[DOI]]

APA
[[Last name, First initial]]. ([[Date Created]]). [[Title]]. [[Publisher]]. Retrieved from [[DOI]]

Chicago
[[Last name, First name]]. "[[Title]]." ([[Date Created]]): [[Publisher]]. [[DOI]]

new citation format to add:
NLM
[[Last name, First initial]]. [[Title]]. [[Publisher]]. [Internet]. [[Date Created]] [cited [[insert today's date]]]. [[DOI]]

NLM example: Pastva, J. Some presentation I gave. DigitalHub. Galter Health Sciences Library & Learning Center [Internet]. 2017 [cited 2018 Oct 15]. doi:10.18131/g3-7k8p-gg29

@kglibrarian

[fenekku]

Just thinking about the future here... What will happen if/when we transfer those to another repository for instance? The DOI ensures uniqueness in some sense, but should the repository (DigitalHub) even be mentioned?

[jpastva]

@fenekku I was approaching the citation style similar to a resource from a database, where typically you do mention the database name. @kglibrarian has found similar examples where Zenodo and the University of Waterloo also included the repository name in their citation. I'm not sure whether this citation is generated on the fly, or if it becomes a part of the object's metadata, but I'm assuming we can change "DigitalHub" if a name change comes later on down the road?

[fenekku]

This is probably a nitpick at the end of the day ;). tldr; sure, we can have the name of the repository in the citations ╮(︶▽︶)╭

The citations are generated for sure, so changing the name in the generated citation is not a problem. Citing established databases makes sense, because they are pretty stable/unchanging (I don't think PubMed is considering a change of name).

We are in this weird scenario where there is a high probability of repository change. So there would be citations referring to an entity that doesn't really exist anymore (at some point). Will this mean we keep track of the older repository to enhance user experience? But then the DOI resolver already does this for us (presumably). Plus, we are humble enough to know that our digital repository will probably be replaced again... So, the value of the name seems more for us: it is probably a vanity thing, which I guess is fine ^_^

[jpastva]

Vanity, indeed!

It is true that if we change the name, there will be old citations out there to our non-existent old repository, but that comes with any name change, and happens all the time with journals. I'd be curious to see if there are any citations out there to DH presently. Perhaps we can give that to @kglibrarian to investigate as a side project :)

[kristiholmes]

Here's an example that came through this week - see Ref 20. :-)
https://www.ahajournals.org/doi/10.1161/JAHA.118.009841

[fenekku]

Ah interesting, the example Kristi gave uses "Galter Health Sciences Library" . Would going that route be less susceptible to change over time?

[jpastva]

The reason I was inclined to remove Galter Library from the citation was to better align with the various citation styles, which, depending on the resource type being cited, infrequently include the publisher name, but always include the "container" of the resource. I also recognize that there's no guarantee that a citation will be preserved exactly as generated from DigitalHub when someone adds it to their references list, so I have low expectations for uniformity in the wild.

[fenekku]

Ok that makes sense.

From discussions here, the Invenio framework uses a CSL library to generate citations. This makes it standard and dependent on what values we provide to it. We will see how that work out in the long run. Thanks!