User talk:Nono314

Jump to navigation Jump to search

About this board

Logo of Wikidata

Welcome to Wikidata, Nono314!

Wikidata is a free knowledge base that you can edit! It can be read and edited by humans and machines alike and you can go to any item page now and add to this ever-growing database!

Need some help getting started? Here are some pages you can familiarize yourself with:

  • Introduction – An introduction to the project.
  • Wikidata tours – Interactive tutorials to show you how Wikidata works.
  • Community portal – The portal for community members.
  • Contents – The main help page for editing and using the site.
  • Project chat – Discussions about the project.
  • Tools – A collection of user-developed JavaScript tools to allow for easier completion of some tasks.

If you have any questions, please ask me on my talk page. If you want to try out editing, you can use the sandbox to try. Once again, welcome, and I hope you quickly feel comfortable here, and become an active editor for Wikidata.

Best regards! Liuxinyu970226 (talk) 00:07, 20 May 2015 (UTC)

Previous discussion was archived at User talk:Nono314/Archive 1 on 2016-01-17.

MediaWiki message delivery (talkcontribs)

Hello Vunj98osrr08knwt,

Really sorry for the inconvenience. This is a gentle note to request that you check your email. We sent you a message titled "The Community Insights survey is coming!". If you have questions, email surveys@wikimedia.org.

You can see my explanation here.


Reply to "We sent you an e-mail"
VIGNERON (talkcontribs)

Bonjour,

Il n'y a pas de référence indiquée sur Q95488754 je te contacte directement. Pourrais-tu vérifier les dimensions de ce portrait ? (les dimensions 101*13 correspondent à un paysage très horizontal ce qui n'est pas le format et l'orientation de l'image )

Nono314 (talkcontribs)

Bonjour @VIGNERON, tu as parfaitement raison ! Il s'agissait d'une coquille désormais corrigée grâce à ta vigilance.

Malheureusement, certains musées ont toujours des systèmes extrêmement pénibles qui ne permettent pas de faire un lien direct vers la fiche d'une œuvre, ce qui complique fortement la remontée à la source pour qui veut vérifier...

Reply to "Dimensions apparemment fausses"

The International Museum Day Wikidata Competition 2020 – Congratulations!

1
Manfred Werner (WMAT) (talkcontribs)
Reply to "The International Museum Day Wikidata Competition 2020 – Congratulations!"
Shonagon (talkcontribs)

Hello Nono314.

J'espère que cela va pour toi.

Pour information, au cas où cela t'intéresse, les vocabulaires Joconde ont été intégrés à Mix'n'Match et une page projet a été créée  : Wikidata:WikiProject Vocabulaires Joconde .

Il y avait des soucis sur les motifs d'identifiants. Ils ont été repris pour s'accorder à tous les cas. Les quelques éditions précédentes ont été si nécessaires corrigées.

Bien à toi


Nono314 (talkcontribs)

Bonjour Shonagon et merci pour la notif.

J'étais tombé sur le catalogue auteurs en début de semaine sur Mix-n-Match et je voulais te contacter à l'occcasion du week end, tu m'as devancé !

Habituellement, je maudis ceux qui créent des catalogues sans définitions explicites, avec juste un terme en dur pour toutes les entrées ;-) Mais il faut reconnaître qu'avec les vocabulaires Joconde, on n'est pas vraiment aidés ! Je me demandais comment tu arrivais à valider les associations : la page projet ci-dessus me permet de comprendre, mais ce n'est effectivement pas une sinécure...

De mon côté, j'avais commencé à préparer l'import à la création de la propriété (c'est là que j'avais découvert le double format d'identifiants !) avec l'idée d'utiliser pour les descriptions le contenu du champ PAUT de Joconde quand il était disponible. Le but étant de permettre une identification aussi rapide que possible, et éventuellement d'extraire les dates de naissance/décès qui permettraient de gérer les cas les plus évidents en auto. À mon sens, c'est un travail qui aurait dû être fait à la source côté ministère, afin de qualifier un minimum leur thésaurus.

Christelle Molinié avait aussi suggéré de réutiliser les associations de JocondeLab, ce qui m'avait fait immédiatement penser à toi, puisque ton bot les exploite déjà pour enrichir et référencer les éléments des œuvres. Tu n'as rien vu de jouable à ce niveau ?

Pour ce qui est des autres vocabulaires, je ne suis pas sûr que Mix-n-Match apporte énormément, la qualité des suggestions étant généralement plus que médiocre sur les noms communs.

Si je peux faire quelque chose, n'hésitez à me solliciter !

et bon courage ;-)

Shonagon (talkcontribs)

Hello Nono314

Vraiment désolé que le Mix'n'Match soit aussi sec mais ce n'était pas possible de faire autrement depuis data.culture. On a également évoquer auprès des interlocutrices du ministère l'éventuelle possibilité d'un export ponctuel plus riche mais ce n'est pas possible, les données n'étant simplement pas dans GINCO.

Oui il eût sans doute été possible de faire mieux avec extraction depuis des scrappings successifs sur POP. Cela aurait été complexe mais pas impossible. J'ai aussi étudié la piste de possibilité d'extraction depuis DBpedia. Malheureusement le périmètre aurait été nettement limité (seulement les auteurs ayant un article Wikipédia en 2014) avec problèmes d'homonymies, libellés différents, changements de page...

Dans tous les cas il aurait fallu un dispositif de validation/correction, ce qu'offre très bien Mix'n'Match avec collaboration possible. Plutôt que de m'égarer sur de longues et incertaines bidouilles dans mon coin, il m'a semblé plus pertinent de publier sur Mix'n'Match le jeu disponible, ce qui est assez facile à faire. Bon, c'est laborieux pour l'alignement mais ça reste jouable. Pour les auteurs, l'utilisation du script de changement de lien de data.culture vers POP est quasiment indispensable pour lever pas mal d'incertitudes.

Comme tu le remarques aussi, on s'est rendu compte que les propositions d'alignement pour les noms communs sont globalement médiocres. Cela pose surtout difficulté pour le vocabulaire Dénomination ; les autres étant soient de volume réduit soit des lieux (où ça fonctionne bien).

De mon côté, je vais me concentrer sur le vocabulaire Auteurs, car j'y vois un intérêt assez énorme côté boulot. En alignant les artistes d'une base avec Wikidata, on peut notamment en déduire l'alignement Joconde et aussi les résultats POP. Par exemple pour Nicolas Pousin avec Q41554, on a http://data.culture.fr/thesaurus/page/ark:/67717/T513-25084/ d'où on peut déduire la page de résultats POP.

Je n'ai rien de spécial à te demander :-). C'était juste pour te tenir au courant, vu que tu connais bien ces questions, as beaucoup contribué dans le domaine et été intervenu dans les discussions.

Bien à toi

Reply to "Vocabulaires Joconde"
Jura1 (talkcontribs)

Hi Nono314

As you are an active user of Mix'n'Match, I was wondering what you think of the approach outlined at Help:Add_main_subject_with_Mix-n-Match, including Wikidata:Property_proposal/MxM_xref.

The idea is to be able to match up items for entries of biographical dictionaries (and other publications with items), e.g. the ones at Help:Import_NBD_from_enwikisource/lists/all_entries. I actually tried it in the test catalogue and it went fairly well.

To move ahead with others, Wikidata:Property_proposal/MxM_xref would be most helpful. --- Jura 11:23, 13 April 2020 (UTC)

Reply to "MxM for article items"
Jarekt (talkcontribs)

Thanks for cleaning up my mess at Oranges (Q66318926). A very helpful next step would be to look at the reference and see where all that info come from ( in this case File:René-Antoine Houasse - Story of Minerva - The Muses Showing Minerva Hippocrene Waters of the River that Brings Out Pegasus, 1696.jpg ) and fix it there as well (someone connected that painting to Oranges (Q66318926) instead of Story of Minerva - The Muses Showing Minerva Hippocrene Waters of the River that Brings Out Pegasus (Q66318923)). Otherwise next time software will notice missing info and copy it again.

Nono314 (talkcontribs)

Thanks, you're right! Actually the root cause was the QS job that created Oranges (Q66318926) with the wrong image (P18) value, but I somehow forgot the step where Multichill's bot sets backlinks in Artwork template, from where you pick up the ball for importing statements. Even if I know the process, it's sometimes hard not to forget one step... Thanks for catching it.

By the way, would it be possible on your side to skip import when there are obvious discrepancies between template on commons and item on wikidata, such as creator or collection? Does your technicall stack for this job allow this? Or maybe this should be handled in artwork module, cataloguing those entries in maintenance categories as you did for discrepancies in authority control when you were working on those? I'm starting to think this would be a nice way to flag inapropriate links made by bots. This case was obvious but I have seen many other cases about different versions of the same painting where it's more difficult to untangle things, and it's not uncommon to have an image used on a copy's item, which leads bots to infer wrong things.

Jarekt (talkcontribs)

Creating maintenance category for all the files with significant discrepancies between commons and wikidata is a good idea, and I tried in several cases. The issue is that such categories fill up with hundreds or thousands of files and it takes a LOT of time to look through them. For example I created

So the bottom line is that I create such maintenance categories but I am usually the only one fixing any issues they highlight.

Nono314 (talkcontribs)

I remember spending quite some time back then on things like creators without home cats or AC discrepancies, probably after you advertised them on Project chat... Nobody usually finds them unless they know about them in the first place. Thanks for the heads up, I didn't know about several from c:Category:Artwork template maintenance. There are many "missing Pxxx" categories in there but things are not always so straightforward, even more with categories often blurring the lines between versions of a painting.

Reply to "{{Q|Q66318926}}"
Keren - WMIL (talkcontribs)

Hi Nono314,

A month ago I ran a batch to create around 1000 items based on the Israeli Museum portal. I see that the next day you've created more items based on the same source. It would have been nice if you had you let me know that you are planning to do this, as this is part of a project and I was preparing another batch which I ran yesterday, only to discover that quite a few items already existed. Also, the items that you created often miss information like labels in Hebrew and English, you've used the registry number and not the inventory number, you didn't put references for the statements, in some cases (like this item) the URL leads to a different painting.

In short, you have have used data of which was clear that I was working on, and now I have more work to sort out what you did and to try and make it consistent with the format of the previous batches I uploaded. This is not the way a collaborative enterprise should work. Please notify me if you plan on using any data from the Israel Museums Portal. Thank you and happy new year. ~~~

Nono314 (talkcontribs)

Hi Keren,

I am not working from the Israel Museums Portal, I'm working on several painters for whom I extend the listing of artworks worldwide (there are several ways to slice work!), primarily out of data listed on museum websites. I probably created about 2 or 3 dozen items for artworks hosted in Israel, not really a scale of action that would disturb anybody else's work, unless they have a very exclusive notion of their field of work. And you can't just request the entire world to stop running for months because you're "working on something"... We all have to cope with moving ground!

As usual when creating additions, I studied existing items to avoid duplicates as well as fit with existing model. The inventory numbers I used where consistent with those listed on the museum's own web site and with those of all existing items previous your mass import from a secondary source. I may of course miss information like additional labels (especially in a language I know nothing about) and you are welcome to add it: that's how wiki projects work, by building on the work of others, not by denying them to work in self-proclaimed exclusive fields. Then, it seems I did one wrong copy paste, that can happen ; you could have just fixed it, but I am happy to do it now you have pointed at it. May I say your work is also not immune from errors? Like claiming this being created in the first century? Yes errors can happen!

You also never cared to answer the message I had left you, so maybe ask yourself whether it's the best way to encourage others to further attempt at "cooperating" with you?

So happy new year to you too, and be prepared: there probably will be other people creating items that may cross your path in the upcoming year, so take it in a positive way! I almost daily find duplicate or questionably filled items that I amend/merge: this enhances the project more than ranting about "more work" ;-)

Keren - WMIL (talkcontribs)

Hi Nono and happy 2020,

I don't expect the whole world to stop running, not do I "work in self-proclaimed exclusive fields", but I do hope for more communication and collaborative spirit instead of each person doing their thing with no consideration to other projects that intersect yours, which causes inconsistencies in the data.

You say: "I may of course miss information like additional labels (especially in a language I know nothing about) and you are welcome to add it: that's how wiki projects work, by building on the work of others," - actually wiki projects also work by working together with other people not just one on top of the other which often is not the most effective way.

Of course errors can happen, and I never suggested that my work is immune from them. And I could have just fixed the error I found, but maybe what caused the error also affected other items, in which case you might want to apply a more global fix. For example, I didn't answer to the message you left me, but I added the missing data by identifying the source of those errors: as the Israeli Museum portal consists of different museum catalogs, there is no consistency in how dimensions are specified (i.e., height and width appear sometimes in the data as width and depth). Also thanks for noticing the date error above - it looks like a bug in openrefine because in the data I have uploaded, P571 s given as 17C.

To sum, I don't think there is just one way in which wiki projects work and I hope that moving forward we can establish a more coordinated, collaborative and effective way to work on our projects.

Reply to "Israel Museums Portal"
Adam Harangozó (talkcontribs)

Hi, I created a new page for collecting sites that could be added to Mix'n'match and I plan to expand it with the ones that already have scrapers by category. Feel free to use, expand. Best, Adam Harangozó (talk) 22:58, 19 October 2019 (UTC)

Reply to "New page for catalogues"

Reminder: Community Insights Survey

1
MediaWiki message delivery (talkcontribs)

RMaung (WMF) 19:53, 20 September 2019 (UTC)

Reply to "Reminder: Community Insights Survey"
MediaWiki message delivery (talkcontribs)

RMaung (WMF) 17:37, 10 September 2019 (UTC)

Reply to "Community Insights Survey"