Open main menu

User talk:Shonagon


Faux numéro d'inventaire ?Edit

Bonjour Shonagon,

J'essaye de faire un petit coup de nettoyage sur quelques peintures du Louvre et je suis tombé là dessus : J'imagine qu'il faut lire RF 2004-18 et non 2000-18 ? C'est d'ailleurs ce qui figure sur --Zolo (talk) 13:45, 29 January 2018 (UTC)

Bonjour Zolo.
Oui en effet ça paraît bizarre pour une œuvre acquise en 2004. RF 2004-18 est assurément bon et c'est RF 2004-17, pour la deuxième œuvre car la notice d'Atlas porte sur les deux. C'est manifestement une coquille, 2000 au lieu de 2004, comme l'atteste d'ailleurs la notice de l'Agence photo de la Rmn :
Du coup, un item a été créé pour le second tableau : Trees in the Prater (Q47516989). Bien à toi --Shonagon (talk) 19:50, 29 January 2018 (UTC)

Bot didn't respect "attributed to"Edit

For Gustav III, King of Sweden, in a Gustavus Adolphus inspired dress (Q18685950) ShonagonBot hasn't respected the "attributed to" in 2015. No idea how many items are affected by this issue. How can we find and fix them? --Marsupium (talk) 19:40, 4 February 2018 (UTC)

Hello Marsupium,
Those edits of ShonagonBot were made in 2014, and there was no policy at this time on attribution. It was the first massive creation of artworks items from Wikimedia Commons. Yes, it is possible to find similar cases :
    ?item wdt:P18 ?img .

Try it!

As you can explore, the issue doesn't concern only those contribution but much more. Fixes are in progress. If you dig artworks items, not only with images files begining with "Attributed to", you will find many creator (P170) statements wich could be qualified with an attribution reservation. And if your are familiar with artworks database, you find it very often. This is art history and doubts about attribution are common, and where Categories for the Description of Works of Art (Q5051819) recommands to use qualifications on statement (4.1.2. CREATOR QUALIFIER). This is a common practice to associate them with an author on database. It allows to reuse and provide data easily to find (the goal of Wikidata), for example all artworks of an artist, including those with attribution reservation or from their studio (distinction is often precious and point of view). Yes, Wikidata should be at the best (on quality and reuse) and that the way I use today, for example I made the attribution qualifications for the paintings of the Louvre. Here we can see, once again, the big interest of the Attribution Qualifier proposal of Vladimir Alexiev that I fully support. Best regards --Shonagon (talk) 19:34, 5 February 2018 (UTC)
To voice my opinion. attributed to (P1773) is a bit of a weird property and I've been considering proposing to get rid of it. With all the other properties like school of (P1780), circle of (P1776) and more the source implies that is someone and we don't know who, but we're confident that person is part of the school, circle, follower, etc. With the attributed to, it's the other way around, we're pretty confident it was made by some person, but it was not that is what signed, but something scholars could agree on. Changing these attributed to cases to sourcing circumstances (P1480) (but not the follower of, etc.) would probably be a good improvement. Multichill (talk) 20:11, 5 February 2018 (UTC)
Thanks Multichill. It seems to be the wisest proposal and I agree with it. Using sourcing circumstances (P1480) for attribution reservation on known and supposed creators but keeping other properties (manner of (P1777),follower of (P1775),...) which, as you say, fills for qualifications on unknow creator but linked to an another known. Best regards --Shonagon (talk) 20:50, 5 February 2018 (UTC)
Thank you both for your responses! Regarding the ontology changes I don't have a strong opinion, but we should definitely discuss it more in depth! --Marsupium (talk) 11:46, 16 February 2018 (UTC)
Please vote: Wikidata:Property_proposal/Attribution_Qualifier --Vladimir Alexiev (talk) 18:22, 25 February 2018 (UTC)

Donald Duck Family Tree (Q630139)Edit

Hi, I noticed you removed drawing (Q93184) from Donald Duck Family Tree (Q630139). You may be right about this, and I've struggled to find the right class to note that this is about a specific work (as shown here). Do you have a suggestion for a better way to do this? Einar Myre (talk) 15:04, 9 February 2018 (UTC)

Hello Einar Myre. Yes, that's it. The item is not an instance of a material drawing (Q93184) but the concept of a family tree (Q189977). I am not familiar with those kinds of items. Here, a list of others, that could help:
   ?item wdt:P31 wd:Q189977 .
   SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
Try it!
For a specific drawing, we need a new item, which be an atwork item, with creator (P170), inception (P571), material used (P186), collection (P195)... and which could have a statment with depicts (P180):Donald Duck Family Tree (Q630139). Best regards. --Shonagon (talk) 11:03, 11 February 2018 (UTC)
Thanks, I'm thinking this actually is the artwork item, so it should maybe rather be an instance of drawing (Q93184) (or something similar) that depicts (P180) a family tree (Q189977). I should clarify, though, that by "specific work" I meant that this is about a specific visual representation of a family tree, rather than a specific object that exists somewhere in the world (so I'm not sure if something like collection (P195) really makes sense here). Would drawing (Q93184) be right for this type of work? Einar Myre (talk) 16:46, 11 February 2018 (UTC)
Hum, obviously there is some confusion with wikipedia articles. The articles in Italian an d Swedish are more about the artwork, the French is much more about the Donald Duck's family that shows the artwork (reason why I removed the instance of drawing (Q93184)). As you say the artwork should be considered at least as important as the concept. Maybe 2 items could be a better solution. I added some statements to the existing item. Yes collection (P195) could be an issue because it is more an illustration (Q178659) than a drawing (Q93184). For the depicts (P180): family tree (Q189977), I added a qualification of (P642):Donald Duck (Q6550). Best regards. --Shonagon (talk) 20:45, 11 February 2018 (UTC)
Looks good to me! I see that illustration (Q178659) isn't in the creative work (Q17537576) tree - should it be a subclass of (P279) visual artwork (Q4502142)? Einar Myre (talk) 14:53, 14 February 2018 (UTC)
Hello Einar Myre. A query on instance of (P31):illustration (Q178659) :
SELECT ?item ?itemLabel ?img
  ?item wdt:P31 wd:Q178659.
  ?item wdt:P18 ?img.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
Try it!
All of those items are visual artworks. They have often a second instance of (P31) like drawing (Q93184) but imho it is not excalty the same thing. As you suggest illustration (Q178659) could be subclass of (P279) of visual artwork (Q4502142), an effective solution claiming that an illustration is itself an artwork. It's not an easy question because there is often debates on what is artwork and what is not. But all of the current items leads to consider that they are. So it was made. Best regards --Shonagon (talk) 00:00, 16 February 2018 (UTC)

Stats bizarresEdit

Bonjour Shonagon,

Je vois que le nombre d'oeuvres sur Crotos progresse bien (notamment je crois grâce au bot de Multichill qui ajoute pas mal d'images Commons). Je ne suis pas sûr en revanche de comprendre comment sont calculées les stats. Par exemple, entre le 4 et le 13 février, il est censé y avoir 6586 oeuvres, mais le nombre total d'oeuvres passe de 119575 - 110759 = 8816 [1].

Bien à toi. Zolo (talk) 17:51, 23 February 2018 (UTC)

Hello Zolo .
Je te prie de m'excuser pour le retard dans ma réponse. Oui il y a une soudaine poussée de nouvelles œuvres et on devrait atteindre dans les mois prochains mois les 100 000 peintures avec image (on en est à 89 000).
Pour l'écart, il y a dans le lot ~2000 œuvres qui avaient été publiées sur Crotos puis retirées car sans images (ça craque un peu sous le volume :-/, même si certes d'un autre côté il y a de quoi se réjouir :-) ; ces œuvres ne sont pas comptées comme nouvelles (il y a une table de la base qui mémorise les publications d'items même retirés ensuite) mais elles rentrent bien dans le nombre total actuel. Cela explique d'ailleurs que le nombre de nouvelles (œuvres avec) images est plus important.
Bien à toi --Shonagon (talk) 21:42, 7 March 2018 (UTC)

Georges AubertEdit

Georges Aubert (Q3102213) & Georges Aubert (Q50413771) is an interesting puzzle. Either it's the same person or two persons completely mixed up. Can you have a look at it? Multichill (talk) 19:14, 20 March 2018 (UTC)

Hello Multichill. Yes indeed an interesting case. I will take a look for it. Best regards --Shonagon (talk) 21:04, 21 March 2018 (UTC)
@Multichill I'm really not sure that they are the same person. The first one, a Swiss painter, , and the second one is a French engraver .Signatures are different: and The difficulty is that they seem to be mixed in some referencials like the Bénézit which gives the date of 1886 for the year of birth of the engraver (instead of 1866 which is coherent with the date of some of his works); or in BnF's catalog whose notices are based in part on the Bénézit for one and for all for the second. Really not an easy case :-/. Best regards --Shonagon (talk) 22:27, 21 March 2018 (UTC)

Share your experience and feedback as a Wikimedian in this global surveyEdit

WMF Surveys, 18:57, 29 March 2018 (UTC)

Reminder: Share your feedback in this Wikimedia surveyEdit

WMF Surveys, 01:40, 13 April 2018 (UTC)

Zimmer exhibition room (Q15206795)Edit

Bonjour Shonagon, j'ai oublié un argument pour notre discussion:, regards--Oursana (talk) 17:21, 26 July 2018 (UTC)

Suite à l'atelier Wikidata de dimancheEdit

Hello Shonagon. Je me rends compte que je ne maîtrise pas encore bien les requêtes SPARQL... Je souhaite par exemple afficher sur une carte tous les éléments wikidata d'un pays (Israël Q801 par exemple) qui contiennent un identifiant de relation (OSM) P402. En m'inspirant de la requête qui affiche tous les hôpitaux sur une carte, j'ai mis ceci :

  ?item wdt:P402;
        wdt:P17 wd:Q801;
        wdt:P625 ?geo .

Try it!

Mais, la requête est évidemment mal formulée. Peux-tu m'aider à trouver la bonne solution ? Amicalement, Mazuritz (talk) 19:28, 12 November 2018 (UTC)

Hello Mazuritz. Il manque juste un complément au premier élément de la recherche (2 crochets pour n'importe quelle valeur / une variable, genre ?osm / ou quelque chose de précis comme wd:Q801) :
  ?item wdt:P402 [];
        wdt:P17 wd:Q801;
        wdt:P625 ?geo .
Try it!
L'idée est que pour tout graphe de recherche on a un sujet / un verbe / un complément (côté jargon, il s'agit du triplet sujet / prédicat / objet du web sémantique).
Le point virgule est utilisé ici ensuite pour indiquer qu'il s'agit du même sujet ; mais ça, tu le sais certainement déjà. Bien à toi --Shonagon (talk) 20:57, 12 November 2018 (UTC)

À propos de Special:Diff/837521904/837523071Edit

Je ne vois pas en quoi c’est un mauvais format pour une date approximative : la source que j’ai indiqué mentionne exactement « 14 apr. J.-C. ». Si cette information est incorrecte, c’est un cas à traiter avec le rang de qualité de l’information, pas une fusion. D’autant que la copie de la source que vous avez faite a pour effet de sourcer une affirmation avec une source qui dit autre chose, ce qui dégrade plus la donnée de mon point de vue. — Ltrlg (talk), 13:46, 19 January 2019 (UTC)

Bonsoir Ltrlg. Oui, tout cela est juste sauf que c'est très vraisemblablement une incorrection d'édition pour une datation identique. L'œuvre est datée 14-23 ap JC par le musée et l'autre déclaration ne la contredit pas. Wikidata est fait pour être réutilisé, et ces doubles déclarations compliquent inutilement. Les dates multiples sont fréquentes notamment suite aux fusions. Il est plus sain de simplifier par exemple en gardant une date du catalogue du musée sur une autre importé de Commons. Les désaccords de datation existent et doivent en effet être séparés et sourcés. Mais ce n'est pas le cas ici. On parle d'une sculpture du premier siècle à propos de laquelle il n'y a pas de datation précise à l'année. Le bon sens est souhaitable. Bien à vous Shonagon (talk) 19:53, 19 January 2019 (UTC)

Mixed upEdit

Hi Shonagon, Thank you for adding data about Dutch Wikipedia articles to Wikidata. I do like to request to make in future some changes, as not everything was added right. In the Netherlands are municipalities the lowest level of government, so we add for located in the administrative territorial entity (P131) only Amsterdam (Q9899). For location (P276) you can insert Amsterdam-Zuid (Q478636), and not the street as we use located on street (P669) for streets. Please keep this in mind when adding new data to items. Thanks! Romaine (talk) 20:22, 10 February 2019 (UTC)

Hello Romaine. Thank you for this clarification; it was not obvious for me indeed. I will be in the future careful by following these recommendations. Those items make a great set of public artworks in the Netherlands. It makes want to discover them in situ and we need good data for that. Best regards --Shonagon (talk) 19:18, 11 February 2019 (UTC)

Image suggestions for paintingsEdit

Hi Shonagon, Wikidata:WikiProject sum of all paintings/Image suggestions/Creator, institution and inventory number match got fixed and it's filling up. Could use a hand! Multichill (talk) 11:47, 16 March 2019 (UTC)

Return to the user page of "Shonagon".