Page MenuHomePhabricator

Visual Editor > Cite Automatic is creating cite journal for URLs
Closed, DeclinedPublicBUG REPORT

Description

Steps to replicate the issue (include links if applicable):

I am editing [[:en:Southport, Queensland]] using Visual Editor and I want to add this URL as a citation

http://digitalbank.goldcoast.qld.gov.au/vital/access/manager/Repository/vital:27316 (an online photo archive)

Cite > Automatic > paste in my URL > Generate

What happens?:

creates a {{cite journal}} instead of {{cite web}}

The Visual Editor user has no way to modify the resultant citation as the citation type is not able to be changed in VE (previously reported problem).

image.png (348×1 px, 116 KB)

What should have happened instead?:

When given a URL as input, I would have expected {{cite web}} to be the default citation type unless there was a positive match to the URLs for known journals. I cannot see anything in the source of the page that suggests it is a journal.

Obviously I am accustomed to Cite > Automatic URL not producing a perfect result for that URL given the vagaries of the information at that webpage, but I haven't seen it confuse a webpage for a journal article before.

Software version (skip for WMF-hosted wikis like Wikipedia):

Other information (browser name/version, screenshots, etc.):

Event Timeline

I can confirm this is recognized as journalArticle by Zotero "COinS" translator.

The Coins translator uses Zotero's parseContextObject which interprets the rft_val_fmt=info%3aofi%2ffmt%3akev%3amtx%3ajournal (rft_val_fmt=info:ofi/fmt:kev:mtx:book) in the title attribute of the page's span.Z3988 element as journalArticle. It should have been rtf_val_fmt=info:ofi/fmt:kev:mtx:dc to be interpreted as webpage (which would have resulted in Citoid returning a Cite web template instead).

I've fixed it (partially) on Web2Cit. Remember you can insert Web2Cit-generated citations either installing the Web2Cit user script, or prepending https://web2cit.toolforge.org/ to the URL you are trying to cite via Citoid. Note that:

  • It's returning webpage as requested, but I wonder whether it should return artwork instead (given that the item refers to a photograph). This would become a Citation template in the English Wikipedia.
  • It's returning nothing for author (which I think is the expected behavior, given that the source says "Creator: Unidentified"). Citoid returns "7.1, Vital Repository".
  • The date is specified as "1900-1909". However, Web2Cit has been designed to only support returning dates in one of yyyy, yyyy-mm, or yyyy-mm-dd formats. Worked around as 1900 (actually, 1900-01-01 because of T309706). I wonder whether Web2Cit should be changed to support additional date formats (T132308 is probably relevant here).
  • Returning "publishedIn" (publicationTitle) and language as well. Citoid isn't returning anything for these.

I've tried it with a couple other webpages from the same website, and it seems to be working OK.

In the long term, I guess that the library's webmasters should be contacted to have the COinS meta tags corrected.

Realistically, we can't do anything about bad metadata on websites upstream, other than the workaround @diegodlh has mentioned, so closing this!