Update generated citations

[fenekku]

Extracted from #52

Harvard Style: For Author first name, only the initial is used, not the whole name. The article title receives single quotation marks (right now there's a double quote at the end and two commas at the beginning.) Reference: http://libraryguides.vu.edu.au/harvard/sample-reference-list

MLA: Both last and first name are fully spelled out. Article title receives double quotes. Reference: https://sites.umuc.edu/library/libhow/mla_examples.cfm#academicjournals

Vancouver: Good, although the date elements at the end need spaces in between, and the extra punctuation at the end and the 1. at the beginning can be deleted. Reference: http://guides.library.uwa.edu.au/c.php?g=324981&p=2178465

[Bioscibrarian]

I just checked the Vancouver style, and, while our version is looking good, it's generally the Vancouver style for a print resource. For electronic resources, Vancouver style guides indicate that the URL should be included and that doi is optional.

Example from the page you linked @fenekku: Watt RG, Rouxel PL. Dental caries, sugars and food policy. Arch Dis Child [Internet]. 2012 [cited 2018 Jan 23];97(9):769-772. Available from: https://adc-bmj-com.ezproxy.library.uwa.edu.au/content/97/9/769

They also recommend to add "[Internet]" and the cited date, but I think this is not necessary. The URL is more relevant, I think. Especially if we want people to use the repository to get cited for their datasets and other objects.

Would it be too hard to add a URL for the record?

[fenekku]

To generate citations, Invenio uses the Citation Syntax Language. We get to configure which formatting variant to use. Like I showed yesterday, there can be multiple variants for a given style (all variations are here: https://github.com/citation-style-language/schema/blob/master/csl-data.json#L12)! I hope we may find a variant that works well among those. If not, we will add our own!

[Bioscibrarian]

Ah yes. I see that there is URL in the fields available--which is not surprising--though before you showed it to us, I had no idea there was such a thing as Citation Syntax Language. I'm familiar with the process of editing fields in a citation style within EndNote. This looks similar and flexible enough that you can add the URL to the Vancouver style.

I think the reason that I want the URL is that it aids in the "findability" part of the FAIR principles for open data. For the same reason, I don't think it would hurt to add the doi, which is optional in Vancouver, but provides one more point of findability and accessibility.

Thanks!

[fenekku]

Oops, what I linked illustrates the different fields that can be used/combined to create a variety of formats. The big list of formats itself is here: https://github.com/citation-style-language/styles/tree/a4bbb9b3b822312638d5141aa3bbfe34518112c8 . My hope is that one of them incorporates the URL for vancouver style. This way we could just re-use it.

[Bioscibrarian]

Yeah, I get it. I took a look at the list, which happens to be truncated at 1000, so the original Vancouver style is cut off -- not sure how to see it. There is an Elsevier Vancouver style and a Nature Vancouver style. Looking more closely at both, it appears that the Nature style only includes URL for pub type of journal article. Elsevier-Vancouver includes URL for "webpage, post-weblog" types, but it does affix the doi for articles. It still seems to be limited when it comes to publication type.

I'll leave the fixing of this to you @fenekku , since you're the expert. It seems like we should be able to get at least the doi in Vancouver, which would help with discovery and access. Thanks!

[kristiholmes]

hi everyone, I realize that I may be popping in the middle of the conversation, but how many citation formats do we need? What else MLA, APA, and NLM. I agree with Pamela about the doi.

what is the relative distribution of the different formatting conventions across the various journals?

thanks for indulging this question...

[fenekku]

The formats we are supporting are APA, Chicago, Harvard, MLA and Vancouver (#52, see stage) . We talked about the other formats on Monday and decided to favour exporting to Endnote for others since there are quite a few variants to say the least. Someone else might know better about the format distribution.

[saragon02]

Formatting distributions across the various journals is an interesting question. I will look into it and ask for input from the DIWG.

[fenekku]

Some more context to understand the difficulty with citation formatting styles:

• Even though citation styles are codified, there are edge cases that lead to interpretation so different formatting exist for each style.
• Journals/Publishers often start with a known citation style and tweak it (e.g. urological-science for some Elsevier journals)
• Citation styles have changed over time: there are multiple editions of them
• Creating our own format is essentially unmaintainable over the long run because of the effort (and standardizing just doesn't work ): forking, learning the citation style language, updating, maintaining.

The first 3 factors partly explain the approx. 1325 different formatting styles available here (translations are another issue but we are exempt from it)

To address these problems:

• We have to be lenient with some little styling issues, because none of these formattings are perfect
• Be more precise and explicit with the citation editions or perhaps use journal references
• Enhance metadata field mapping from our model to CSL
• Offload issue to citation managers by #357

Citation formatting styles tried so far:

In italics the ones we use now (the others are the ones we tried so far)

APA: apa Chicago: chicago-fullnote-bibliography Harvard: harvard7de, elsevier-harvard MLA: modern-language-association-8th-edition, modern-language-association, modern-language-association-with-url Vancouver: vancouver, elsevier-vancouver, springer-vancouver, vancouver-author-date