Wikidata:Property proposal/Erudit article ID

Erudit article ID

Originally proposed at Wikidata:Property proposal/Authority control

Done: Erudit article ID (P9108) (Talk and documentation)

Description	unique identifier for an article in Érudit
Represents	Érudit (Q3591464)
Data type	External identifier
Domain	scholarly article (Q13442814)
Allowed values	[0-9]{5,10}ar
Example 1	[Men and the paternal function in Caribbean families.]. (Q52585003) → 014516ar
Example 2	La gestion de la différence culturelle par les éducateurs sociaux dans leur travail au quotidien (Q60884238) → 1036696ar
Example 3	Investir les journaux à l’ère industrielle. Fatalité ou opportunisme ? L’exemple de Berlioz (1830-1838) (Q104524554) → 1069468ar
Example 4	De l’hypothèse de la documentation comme technique de résistance et du wiki comme objet de ces résistances (Q104081705) → 1067414ar
Planned use	Backfilling the property for all Erudit articles on WD (~9300)
Expected completeness	eventually complete (Q21873974)
Formatter URL	https://id.erudit.org/iderudit/$1
See also	Érudit journal ID (P4722), Persée article ID (P8758), PubMed publication ID (P698)

Motivation

Français : Érudit (Q3591464) est une base de données en ligne d'articles (résumé et corps de l'article) de publications académiques de langue française. Nous avons déjà une propriété associée Érudit journal ID (P4722) pour les publications. Nous avons un nombre significatif d'articles de ces publications (dans les 9300) ; ils sont tous accessibles sur le site Érudit.org et chaque article est identifiable par son identifiant unique.

L'identifiant Érudit est visible sur chaque page d'articles. Par exemple, l'article Le bien-être économique et la santé des personnes âgées au Québec (Q60430104) accessible à cette URL a l'identifiant "010225ar". Cet identifiant unique est utilisé par le DOI (P356) pour construire son propre identifiant, donc l'import de la propriété n'est sera plus qu'aisé.

Plusieurs raisons pour l'existence de cette propriété (indépendamment du DOI):

DOI.org est un service tiers qui a ses propres problèmes (erreur 404 sur ces exemples 1 2 alors que 1060109ar ou 1060115ar sont accessibles sans problème) ;
Certains articles n'ont pas de DOI: Exemple ;
cet identifiant peut être utilisé pour référencer des propriétés sur un item WD de la même manière que PubMed publication ID (P698) est utilisé ici par exemple.

Si la propriété est approuvée, je ferais tourner un script pour renseigner cette propriété pour chaque article. --Deansfa (talk) 18:46, 28 December 2020 (UTC)[reply]

English: Érudit (Q3591464) is an online database with papers' abstract and content of French-language academic journals. We already have the associated Érudit journal ID (P4722) for journals. Wikidata currently has a significant amount of these articles (a little more than 9300) ; they're all available on the Erudit platform and are identified by a unique "Erudit ID".

The Erudit article ID is visible on each article's page. For example, the article Le bien-être économique et la santé des personnes âgées au Québec (Q60430104) accessible at this URL has the ID "010225ar". It's actually part of the DOI (P356), which will make the backfill easier to perform.

Several reasons for this identifier to exist (independently from the DOI):

DOI.org is an external service that has issues: 404 errors when trying to access Erudit articles (example 1, 2), while direct access to Erudit website work perfectly (1060109ar or 1060115ar).
Some articles don't show a DOI: Example
this identifier can be used to add references to properties the same way PubMed publication ID (P698) is used here for example

If this property is approved, I will run a script to backfill the property for each article (it will parse the DOI (P356) and format the value correctly). --Deansfa (talk) 18:46, 28 December 2020 (UTC)[reply]

Notified participants of WikiProject Authority control

Notified participants of WikiProject France

Notified participants of WikiProject Canada --Deansfa (talk) 22:56, 28 December 2020 (UTC)[reply]

Discussion

Support (as proposer) --Deansfa (talk) 22:18, 28 December 2020 (UTC)[reply]
Oppose unnecessary duplication since identifier is a component of the DOI. Simon Cobb (User:Sic19 ; talk page) 22:20, 28 December 2020 (UTC)[reply]
1) The DOI is a third-party website which has its own issues, I run into 404 errors when trying to reach Erudit articles (error here but works well through Erudit ID). I just created this page to keep track of the issue. 2) the identifier is a transformed version of a component of the DOI. the DOI is uppercase, this id is lowercase. 3) As explained in the (updated) presentation, the property can be used to add references to item's properties. --Deansfa (talk) 22:39, 28 December 2020 (UTC)[reply]

The DOI is not a third-party website, it is a persistent identifier that resolves to the host url of a digital object - if you report the DOI as broken via the form, the registation agency will prompt the host to update the url to make DOI work again. Simon Cobb (User:Sic19 ; talk page) 08:14, 29 December 2020 (UTC)[reply]

By third-party I was referring to the fact that doi.org is acting as an external redirect service between the requester and erudit.org. I never denied the fact that DOI is immutable and persistent. So if I follow well, when a non-English speaker is on the page of an WD Erudit article item, and the DOI doesn't work, he has to fill a form in English and wait that the third-party service is fixed? I'm sorry but Wikidata is multilingual, and we shouldn't assume that everyone is confortable with English. BTW this Erudit article ID is useful beyond accessing article without filling an English form. --Deansfa (talk) 13:24, 29 December 2020 (UTC)[reply]

The point I am trying to make is related to the cause of the problem rather than the solution; a likely reason that the DOI does not resolve is because the URL of the article has changed but the host has not sent the details to the registration agency. Apologies for the misunderstanding, I was not suggesting that it is the responsibility of Wikidata editors to fill out the form. Simon Cobb (User:Sic19 ; talk page) 21:19, 29 December 2020 (UTC)[reply]

Support - each article has a unique ID - direct access through this ID should be allowed easily --Hsarrazin (talk) 14:12, 29 December 2020 (UTC)[reply]
Support Would be usefull. I use time to time article on Erudit on frwiki. It will help a create item of the article. --Fralambert (talk) 14:38, 29 December 2020 (UTC)[reply]
Support Wikidata is going to be more and more a major hub of identifiers, and that is relevant also because if one doesn't work (for its own issues), others are available. This seems to be an example. --Carlobia (talk) 15:52, 29 December 2020 (UTC)[reply]
Support Good base for Wp. --Yanik ^B 22:02, 29 December 2020 (UTC)[reply]
Support Useful (as far as I understand). --Benoît (discussion) 09:20, 31 December 2020 (UTC)[reply]
Support --Jneubert (talk) 07:00, 2 January 2021 (UTC)[reply]
Oppose this simply duplicates information as Sic19 pointed out, it does not provide any redundancy as one identifier is a subset of the other, therefore this just stores information twice leading to maintenance problems. --Hannes Röst (talk) 02:07, 5 January 2021 (UTC)[reply]
Strong support Sic19 and Hannes Röst Any organization accessing a DOI identifier is not prevented from accessing WD for their own identifiers. If we can't register Erudit IDs, why would we accept DNB IDs? Examples are presented where DOIs are not necessarily present. You confuse perennial identifiers with the rest of the identifiers: not all ark, URN or DOI have the majority of identifiers. Cordially. —Eihel (talk) 18:02, 13 January 2021 (UTC)[reply]
- Comment my understanding was that the lack of a DOI for the presented examples represents a technical glitch and not a long term problem. It seemed to me that this property was solely proposed as a short term workaround for a technical limitation with no long term benefit. Of course I am willing to be corrected if that is not the case and there is indeed a subset of Erudit IDs that are not assigned DOIs on purpose and we need the Erudit ID separately to link to those. However, if that is not the case then I would still argue to take the long term view instead of the short term. Yes in the short term, this will make the hyperlinks work but in the long term it introduces duplicate information. --Hannes Röst (talk) 01:24, 15 January 2021 (UTC)[reply]
  - We're using WD articles information on the French Wikipedia through the template {{Modèle:Bibliographie}} link. If I understand, when a French speaker wants to click on the link of the article in the French Wikipedia, the person just have the option of the DOI, which is broken sometimes as I showed earlier, or just don't exist. Erudit is a meta platform by itself, and French Wikipedians should have the option to access the article through the DOI or through other websites. What does a French speaker has to do in the current situation? Learn English, and fill a form in English for the DOI to work. Does the all point of the language of the Wikipedias is for the knowledge to be accessible by everyone? --Deansfa (talk) 23:54, 15 January 2021 (UTC)[reply]
    - @Deansfa: I dont think that is how it should work, I dont think there should be a need for anybody to learn English to fill out a form - I think that is a misunderstanding. The idea here would be that if the DOIs are not working then the people at Erudit need to fix it and since they are the ones that manage the DOI allocations -- they will be able to fix the DOIs that are not working. They should be able to update the mapping of DOI -> URL directly. I hope / assume they have a way of contacting them in French since they produce a French website. Again, I am only trying to help and trying to pointing out a potential problem here, but I also see that most people are in favour of the new property. --Hannes Röst (talk) 15:14, 19 January 2021 (UTC)[reply]
      - The goal of Wikidata is to be used outside of this current website. If we want to get the EruditId of an article for a given purpose, we should be able to do so, without parsing a third party ID (the DOI one). the French Wikipedia Template "Bibliographie" I mentioned earlier was a good illustration of it. I'm stepping out of the conversation. --Deansfa (talk) 15:47, 19 January 2021 (UTC)[reply]
Comment maybe a compromise solution could be to store two formatter URLs: https://id.erudit.org/iderudit/$1 and https://doi.org/10.7202/$1 ? --Hannes Röst (talk) 15:14, 19 January 2021 (UTC)[reply]