User talk:Marsupium

Q318117 - Geburtsdatum 1383-1384

Hallo Marsurpium, wie schön das Du hier mitarbeitest. Sachlich hat sich hier meine Frage inzwischen verändert, aber es bleibt ja grundsätzlich interessant, ob auch wie bei template:other date Zeitspannen als Geburtsdatum berücksichtigt werden. zB: 1383/1384. Gebe ich dann beide Daten ein, s. urprüngl. Masolino. Eigentlich ist das doch eher kontroversen verschieden belegbaren Daten vorbehalten. Also das was im Template:Artwork unter template:other date hinkommt, muss doch hier auch seinen Platz haben. Und die ähnliche Frage zu ca. ..(grundsätzlich). Die Alte Pinakothek benutzt hier glatte Zahlen. Ist es gewünscht, dies als Beleg anzugeben? Gruß--Oursana (talk) 01:59, 17 November 2013 (UTC)

Hallo Oursana, ich bin gleichfalls sehr erfreut Deine Signatur zu lesen! Ich bitte für die arge Verzögerung um Entschuldigung, inzwischen habe ich ein paar Überlegungen zusammengesucht:

• Mit der Modellierung von nicht exakten (Zeit-)Daten habe ich mich noch kaum beschäftigt und ich fürchte die Vereinbarungen darüber auf Wikidata sind insgesamt noch nicht sehr weit, so kann ich leider kaum befriedigend antworten.

• Der Beschreibung des Time Datentyp unter m:Wikidata/Data model#Dates and times lässt sich entnehmen, dass die Präsizion angegeben werden kann als „0 - billion years, 1 - hundred million years, ..., 6 - millenia, 7 - century, 8 - decade, 9 - year, 10 - month, 11 - day, 12 - hour, 13 - minute, 14 - second.“ In diesem Fall hast Du dies nun anders gelöst, aber diese Präzisionsangaben reichen natürlich nicht aus.

Besten Dank für Deine Antwort, ich bin ja hier jetzt schon geübter und habe auf Deine Ausführungen hin gesehen, dass bei Datum:erweiterte Anpassungen decade und century angegeben werden kann, immerhin Gruß und einen schönen 1. Advent wünsche ich Dir weiterhin--Oursana (talk) 15:00, 1 December 2013 (UTC)

• Ich meine, dass auf jeden Fall eine Möglichkeit gefunden werden muss, die Daten so zu modellieren, dass keine Informationen verloren gehen, also auch eine Möglichkeit, dass „wie bei template:other date Zeitspannen als Geburtsdatum berücksichtigt werden“, die Informationen müssen hier auch Platz haben, da stimme ich Dir vollkommen zu. Deinen ersten Ansatz halte ich tatsächlich auch für unglücklich. Er wurde für <Edward the Confessor (Q130005)> date of birth (P569) gemäß der Diskussion unter Property talk:P569#More than one date so umgesetzt, nunja.
• Wir müssen auch Fälle, in denen wir nur einen terminus post quem (Q1102542) haben, usw. irgendwie sinnvoll modellieren. Ich weiß leider auch noch nicht, wie genau wir das machen sollen.
• Ein passender Ort für Vereinbarung könnte Help:Modeling#Time & dates einmal werden.

• Es lassen sich sicher bessere Quellen auftreiben als http://www.pinakothek.de/en/masolino-da-panicale. Eine wissenschaftliche Quelle im Sinne von de:WP:BLG ist es wohl kaum, aber besser als nichts denke ich, zumal sie sich immerhin schnell nachsehen lässt.

Beste Grüße, --Marsupium (talk) 17:56, 29 November 2013 (UTC)

Nachtrag: Mehr weiß die Runde, die unter Wikidata:Bistro#« né vers 1132 » ou « né entre 1132 et 1150 » diskutiert. Wir sollten anregen auf Englisch und an anderen Orten weiter zu diskutieren, vielleicht auf der Help:Modeling#Time & dates zugehörigen Diskussion. --Marsupium (talk) 04:18, 30 November 2013 (UTC)

Iconclass

http://www.iconclass.nl/home habe ich gerade nl:Gebruiker:Vincent Steenberg gefunden. Gruß --Oursana (talk) 18:17, 16 December 2013 (UTC)

Ich verstehe nicht ganz, was Du meinst? Ist das eine Anmerkungen zu meinen Äußerungen unter Wikidata:WikiProject Visual arts/Questions#Massacre of the Innocents by Rubens (Q1247103)? Beste Grüße, --Marsupium (talk) 18:40, 16 December 2013 (UTC)

Genau und Jane hatte dort auch die Fundstelle von RKD angegeben. Gruß --Oursana (talk) 19:14, 17 December 2013 (UTC)

Copy of en label to es label

In 99.9% of cases the editions are correct, also I have the consensus of Wikipedia in Spanish to make these changes. If there are no more problems the bot will continue. --Kizar (talk) 23:20, 5 January 2014 (UTC)

OK, thank you for your answer, --Marsupium (talk) 08:39, 6 January 2014 (UTC)

Wikidata:Database reports/Constraint violations/P180#"Type Werk (Q386724)" violations

especially portrait (Q134307) see Property talk:P180#changing/permitting constraint, Wikidata:WikiProject Visual arts/Questions#Kindermord in Betlehem (Q15676570) LG--Oursana (talk) 17:45, 19 July 2014 (UTC)

Oh, really, there are some … Thanks for prodding to the two discussions! --Marsupium (talk) 20:00, 19 July 2014 (UTC), corrected 20:18, 19 July 2014 (UTC)

painting (Q3305213)

Why did you remove the subclass of (P279) work of art (Q838948)? That broke the chain to work (Q386724). Also visual artwork (Q4502142) looks a bit weird/incomplete. Multichill (talk) 19:51, 28 July 2014 (UTC)

I removed it based on the discussion Wikidata talk:WikiProject Visual arts/Item structure#Root item I had with Zolo. And I think that work of art (Q838948) is still a messy item that should not be used as a superclass, far from being well-defined as well as work (Q386724). But I am happy to talk about these classes. --Marsupium (talk) 20:33, 28 July 2014 (UTC)

I'd say it is ok to keep it. It makes some intuitive sense to look for subclasses of artworks and get painting, sculpture etc. aven though we should not not take work of art (Q838948) too seriously. Possibly, we should give "subclass of artwork" a normal rank and subclass of artificial physical object (Q15222213) a "preferred" rank. --Zolo (talk) 21:03, 28 July 2014 (UTC)

I am afraid that a handling like that – intuition and ranks for subclass of (P279)-statements – will lead to more inconsitencies than one clear line. However, it might be OK. Do you have any concrete use in mind for the classification under discussion, Multichill? Perhaps we can find a third perfect solution … --Marsupium (talk) 21:23, 28 July 2014 (UTC)

Strange removal

Hi,

Why did you remove made from material (P186) on oil painting (Q174705) on some items about paintings like The Tiger Hunt (Q544488), Wolf Hunting (Q17484684) and The Museum of Rennes in around 1900 (Q17484654) ? Did I made a mistake or did you ?

Cdlt, VIGNERON (talk) 15:34, 22 September 2014 (UTC)

Hey! The description of oil painting (Q174705) is "process of painting with pigments that are bound with a medium of drying oil". That is not in the value range of made from material (P186). There is even a type constraint on material (Q214609) subclasses at Property talk:P186 – though that might not really be realised. The common way to express what apparently was intended is made from material (P186) <oil paint (Q296955)>, e.g. at Mona Lisa (Q12418). Aussi cdlt, --Marsupium (talk) 17:49, 22 September 2014 (UTC)

Oo I see, it's my mistake, in French the two items have the same name ; I didn’t pay enough attention. Sorry.

Cdlt, VIGNERON (talk) 13:45, 25 September 2014 (UTC)

Italian Evening Landscape (Q20017804)

Hi Marsupium . I come again with another way on those painting and I see that are two different version one is at Alte Pinakothek (Q154568) and the the other Schleißheim palace complex (Q680166). There are no duplicates and this is a bad merging. Could you split them again? Thanks Shonagon (talk) 18:51, 13 June 2015 (UTC)

That is why I oppose to add artworks without inventory number (P217). I do not see evidence that they are two artworks. Visually the files are very similar. The collections of Alte Pinakothek (Q154568) and Schleißheim palace complex (Q680166) are both operated by Bavarian State Painting Collections (Q812285) and exchange their paintings. So the data of the Yorck Project and the Web Gallery of Art reflect just a changeable state each. Hence I think they are both photographies of inv. no. 266. --Marsupium (talk) 13:53, 14 June 2015 (UTC)

Ammonia / ammonium hydroxide

Hallo Marsupium, thank you for checking changes on Wikipedia. E.g. you corrected a mistake of me on Fulbe Q202575 vs. AAT.

Nevertheless, i can not agree with your revert on Q421888 (de:Ammoniumhydroxid en:Ammonium hydroxide fr:Ammoniaque nl:Ammonia), which I, through Mix n Match, connected with the item "ammonia water" of AAT. Your explanation was "incorrect", which is eh... not correct... Some confusion can be explained by the fact that the wikidata item and most (but not all, see: fr, pl, nl, sk, io)) linked wikipedia articles seem to focus on a chemical compound NH₄OH or ammonium hydroxide. However, also the en- and de-version state very clearly that NH₄OH does not really exist. --Paulbe (talk) 17:03, 27 July 2015 (UTC)

I tried to be clear and concise and correct, which turned out to be quite challenging. I spent much more time in writing my above reaction, then i expected. And it is, though correct, not complete. If you need more information, please ask. Yours, --Paulbe (talk) 17:23, 27 July 2015 (UTC)

Thank you for your work on the – really not that trivial – mapping to AAT and for your post and sorry for my overhasty revert. I found ammonia solution (Q421888) on Wikidata:Database reports/Constraint violations/P1014#"Single value" violations and removed ammonia water (300015156) before reading the note of ammonium hydroxide (inorganic material) (300266781) which points out that it refers also to the solution. So do you also think that it is a mistake of the AAT? In general the qualitiy of the definitions of the AAT is quite high so that I thought that the two entries would rather require two items on Wikidata. But here AAT seems to have two entries for the same thing. As it looks like a mistake to me I wrote to the managing editor of the Getty Vocabulary Program and hope to get an answer at the end of August after holidays. Yours, --Marsupium (talk) 15:49, 28 July 2015 (UTC)

Hi @Paulbe: I got this answer: „They are currently linked through associative relationships. "ammonium hydroxide" is "constituent of" "ammonia water." Ammonium hydroxide would be an alkali [NH4+][OH−] that ideally should be a separate record from the "solvent" made from it. However, I see your point, since it cannot exist except in solution. Conceptually they may be separate, even though in physical reality they are almost always or always the same thing. If I cannot find warrant for them as separate entities, we will merge them. In any case, the SNs need to make the distinction clearer.“ And indeed the distinction between the AAT records has been clarified. I think we could perhaps create a separate item for the hypothetical salt. I asked how to handle this at Wikidata talk:WikiProject Chemistry#Items for hypothetical compounds? Cheers, --Marsupium (talk) 14:25, 17 April 2016 (UTC)

Reminder for myself: Discussion now on Wikidata talk:WikiProject Chemistry/Archive/2016#Items for hypothetical compounds?. To prevent those edits a new item has still to be created … :) --Marsupium (talk) 15:46, 25 April 2017 (UTC)

collection

Hi Marsupium, I see you're adding new paintings like Q22116937. Great! Could you please add the collection (P195) ([1]) to these new paintings? I noticed because you're new items popped up at User:Multichill/Paintings without collection. Thank you, Multichill (talk) 20:11, 19 January 2016 (UTC)

Oh, of course, sorry! It was my stupid mistake. They belong all to Bavarian State Painting Collections (Q812285). I'll try to fix it soon! --Marsupium (talk) 21:47, 19 January 2016 (UTC)

Thanks! No worries. I used this query to find the items that need the collection. Multichill (talk) 22:17, 19 January 2016 (UTC)

Thank you! Done! --Marsupium (talk) 22:38, 19 January 2016 (UTC)

Hast Du

dafür einen Beleg? Gruß--Oursana (talk) 20:47, 5 November 2016 (UTC)

Oh! Das ist eine sehr gute Frage! Ich habe auch gerade Gegenteiliges gefunden. Ich fürchte ich habe hier eine Bringschuld, der ich baldmöglichst ausführlicher – denn das scheint notwendig zu sein – nachzukommen versuchen werde! Ich hoffe zwei Änderungen in frwiki und augsburgwiki.de sind 2014 nicht falsch und durch meine Schlampigkeit zustande gekommen! --Marsupium (talk) 15:16, 6 November 2016 (UTC)

Ich habe es gelöscht. Ich kann mich nicht mehr erinnern, wo ich es gelesen habe, ob ich es womöglich verwechselt habe. Wenn ich es doch finden sollte, füge ich es wieder hinzu. Ich habe sogar einen gegenteiligen Beleg gefunden: „In Anwesenheit des Stadtvogtes Jörg Ott und des Stadtschreibes Dr. Konrad Peutinger fand in dessen Hause am 20. (21.?) Dazember 1497 das Verlöbnis des Malers mit Anna, der etwa 20jährigen Tochter des Georg und der Barbara Allerlay statt.“ (Hannelore Müller: Die Malerfamilie Burgkmair, in: Bilder aus dem Bayerischen Schwaben, S. 52). --Marsupium (talk) 14:02, 7 December 2016 (UTC)

und auch hier--Oursana (talk) 17:33, 25 April 2017 (UTC)

Ja, ich habe keine Ahnung, wie ich darauf gekommen bin :o --Marsupium (talk) 07:40, 26 April 2017 (UTC)

champagne stemware

Hi! champagne stemware (Q1635923) in enwiki is "Champagne glass", and the description exactly matches aat:300043247 Champagne glass, which says the same (shallow saucer or flute). I've added that AAT ID with normal rank.

Q1635923 in at least 4 other languages is described as "flute", so that matches aat:300197100: I've added that as deprecated rank. So the problem is that the enwiki article doesn't exactly match the articles in those 4 other languages.

Please don't delete the AAT info: when someone gets around to split the WD item, he'll know to split the AAT IDs. Do you agree? --Vladimir Alexiev (talk) 15:09, 29 April 2017 (UTC)

I think neither is really correct and would prefer not to add a Art & Architecture Thesaurus ID (P1014) at all. Now it shows up at Wikidata:Database reports/Constraint violations/P1014#"Single value" violations. I'd rather try to shorten that list than enlarging it. It is quite easy to add AAT IDs that almost fit, it is a lot of work to do the sitelink and sub-/superclass sorting on Wikidata that improve the Wikidata structure and end up in better matching with AAT.

In this case I would better do the item splitting than adding two Art & Architecture Thesaurus ID (P1014) values, but I guess we have no general guideline on this and it is kind of up to each of us … :-) --Marsupium (talk) 15:28, 6 May 2017 (UTC)

Bavarian State Painting Collections

I see you're working on the Bavarian State Painting Collections. I'll have a robot import the missing paintings. Multichill (talk) 10:18, 11 May 2017 (UTC)

Cool!!! Do you think it is doable to import all fields from there? You're gonna use https://www.sammlung.pinakothek.de? I'd like to help with the mapping that is necessary at least for the creator, material and location fields. --Marsupium (talk) 11:05, 11 May 2017 (UTC)

I fired up the bot for a first pass. See for example Q29896861. Haven't done location (P276) yet, that can be done in another pass. Creator will be added by another bot tonight based on the name. For material I currently only support oil on canvas. Total is 18400 paintings so this might take a while to import. Once we have more works online, I'll probably split it up like the Rijksmuseum. Multichill (talk) 14:45, 11 May 2017 (UTC)

Descriptions and creator (P170)

@Multichill: Great work! However, perhaps hold on and stop the bot for a moment!

On: creator (P170): There are quite a lot of tricky cases like "Abraham van Diepenbeeck (zugeschrieben)"[2] that have to be converted like this.

You better should not set the labels/descriptions to "Abraham bewirtet die drei Engel (zugeschrieben)"@de and "Gemälde von Abraham van Diepenbeeck"@de but "Abraham bewirtet die drei Engel"@de and "Gemälde von Abraham van Diepenbeeck (zugeschrieben)"@de and "painting by Abraham van Diepenbeeck (attributed)"@en and likewise for nl!

And there are even worse cases with two artists in one field! The database is not the best formalized one …

--Marsupium (talk) 15:34, 11 May 2017 (UTC)

Similar cases:

• "Kopie nach", e.g. Q29897230
• "(?), e.g. Q29897206

--Marsupium (talk) 15:39, 11 May 2017 (UTC)

I noticed. I don't see a lot of occurrences yet and anonymous works like this generally need a bit of manual attention anyway (based on my previous imports). I'll leave it running for now and keep an eye on it. I wanted to make a sparql query based on [3] that would get all the items that have a label like "<something>(<nach/kopie/etc>)", but don't have it yet,. Can't seem to get the REGEX to work. You an idea how to fix it? Multichill (talk) 16:00, 11 May 2017 (UTC)

At least 9 now: :-)

• "zugeschrieben": Evening Landscape (Q29896935), Q29896973 Q29897323, Q29897310
• "Kopie nach": Q29897230
• "?": Q29897175, Q29897206, Q29897280, Q29897283

The (hopefully complete) list of those cases I collected from the PDF version of the inventory before https://www.sammlung.pinakothek.de existed and a mapping for Wikidata:

BStGS string	Count	Property	Note
'Art des'	23	manner of (P1777)
'Kopie nach'	181	after a work by (P1877)
'Nach'	3	after a work by (P1877)
'Nachahmer'	51	manner of (P1777)
'Nachfolger'	65	follower of (P1775)
'Schule'	36	school of (P1780)
'Umkreis'	50	circle of (P1776)
'vermutlich'	12	possible creator (P1779)
'Werkstatt'	60	workshop of (P1774)
'zugeschrieben'	95	P1773 (P1773)
'?'	<111	possible creator (P1779)
'Anonymer Meister seiner Werkstatt'	25	workshop of (P1774)
'Werkstatt?'	2		different solution necessary, perhaps workshop of (P1774) with sourcing circumstances (P1480) presumably (Q18122778)
'Replik'	1	after a work by (P1877)
'Werkstattkopie'	20	workshop of (P1774), after a work by (P1877)

For copypasting into code:

(from this script (Github)):

'(Art des)', '(Kopie nach)', '(Nach)', '(Nachahmer)', '(Nachfolger)', '(Schule)', '(Umkreis)', '(vermutlich)', '(Werkstatt)', '(zugeschrieben)', '(?)', '(Anonymer Meister seiner Werkstatt)', '(Werkstatt?)', '(Replik)', '(Werkstattkopie)'

(from this script (Github)):

       creatorqualifierdict = {'Art des': ['P1777'],
                               'Kopie nach': ['P1877'],
                               'Nach': ['P1877'],
                               'Nachahmer': ['P1777'],
                               'Nachfolger': ['P1775'],
                               'Schule': ['P1780'],
                               'Umkreis': ['P1776'],
                               'vermutlich': ['P1779'],
                               'Werkstatt': ['P1774'],
                               'zugeschrieben': ['P1773'],
                               '?': ['P1779'],
                               'Anonymer Meister seiner Werkstatt': ['P1774'],
                               'Werkstatt?': ['ERROR NO MAPPING: Werkstatt?'],
                               # FIXME: Wikidata ontology is not sufficient
                               'Replik': ['P1877'],
                               'Werkstattkopie': ['P1774', 'P1877']}

So there has to be put some more work into the SPARQL query or queries … :-) --Marsupium (talk) 16:06, 11 May 2017 (UTC), 16:18, 11 May 2017 (UTC), 16:18, 11 May 2017 (UTC)

That's a nice list! If you parsed the whole PDF: How many cases did you encounter? I put the bot code on Github. Multichill (talk) 16:19, 11 May 2017 (UTC)

You might like Wikidata:Property proposal/Pinakothek artist ID. Multichill (talk) 16:31, 11 May 2017 (UTC)

Thanks for the code and the proposal and everything! I have never counted, but should be easy with the product of my parsing efforts: https://github.com/jjmhtp/art-data/blob/master/data/bstgs_inventory.csv --Marsupium (talk) 17:06, 11 May 2017 (UTC)

Now I have done the counting, results in the table above. --Marsupium (talk) 17:21, 11 May 2017 (UTC)

instance of (P31) and material used (P186)

Inferring this is not always correct. E.g. this work is a drawing. --Marsupium (talk) 17:53, 11 May 2017 (UTC)

If it's a drawing, why does the museum classify it as a painting? Multichill (talk) 18:06, 11 May 2017 (UTC)

Good question! I'm quite sure theThe database is not really high quality … I think the way to filter out as many non-paintings as possible is to parse/match the material field, here for example "Papier" is suspicious. --Marsupium (talk) 18:11, 11 May 2017 (UTC)

Sorry, I'm wrong probably. I came across this case a while ago. It is just that there is "Zeichnung" meaning "drawing" in its title. Perhaps it is a painting though, perhaps it is actually a drawing. I have investigated the issue further now and it seems there are not that many of those things and it is probably not possible to find them somehow systematically anyway. Sorry for the consideration! --Marsupium (talk) 18:35, 11 May 2017 (UTC)

Just a last question for today: Could you somehow create a list of different entries in the material and location fields in the next run? I'd then create a mapping table so that the information can be used here. --Marsupium (talk) 18:43, 11 May 2017 (UTC)

Inventory numbers

If you look at Wikidata:WikiProject sum of all paintings/Collection/Bavarian State Painting Collections at the inventory numbers, you'll notice several types. Do you know the meanings of them? For example "L <id>" is a loan. Multichill (talk) 19:31, 16 May 2017 (UTC)

The "most important groups" are explained under ABOUT THE ACCESSION NUMBERS. --Marsupium (talk) 12:35, 17 May 2017 (UTC)

location

• BStGS locations:

  # BStGS locations
  SELECT ?item ?itemLabel WHERE {
  {
    ?item wdt:P361 wd:Q812285 . # parts of BStGS
  } UNION {
    ?item wdt:P361 ?sub1 . # parts of
    ?sub1 wdt:P361 wd:Q812285 . # parts of BStGS
  }
    SERVICE wikibase:label { bd:serviceParam wikibase:language "en,de,fr,it,es,pt,ca" . }
  }
  

  Try it!

I mapped some of the locations. The rest can be done later. Multichill (talk) 19:27, 16 May 2017 (UTC)

Great, thank you! --Marsupium (talk) 23:45, 16 May 2017 (UTC)

How did you do the mapping? This – like many other location statements – seems to be wrong unfortunately. --Marsupium (talk) 10:25, 22 May 2017 (UTC)

@Multichill: Great that you have found a way to get over the page 200 threshold of https://www.sammlung.pinakothek.de/api/ and thank you very much for all your work! Could you take out the code that generates the location (P276):Alte Pinakothek (Q154568) statements? They are wrong in most cases … Thank you! --Marsupium (talk) 08:46, 26 May 2017 (UTC)

Or are you using the information given under "Stock"? --Marsupium (talk) 08:53, 26 May 2017 (UTC)

What is wrong with the stock? What would be the correct one according to you for Q29965679? Multichill (talk) 20:10, 26 May 2017 (UTC)

Sorry! I thought about it another time, I think it's fine. I'm not sure if the "Stock"/"Bestand" really indicates the location, or e.g. a storage inside or outside Alte Pinakothek building that just belongs to the Alte Pinakothek sub-organization of BStGS so to say. But it is good how you imported it I think at is valuable information that a painting is in either perhaps even if we don't know perfectly and cannot view the painting in whatever storage. The "Displayed" field is more precise, often it reads "Not on display" though – as in this case. I think here it is perhaps the best solution to use both the "Displayed" and the "Stock" field and take for the value of location (P276) an item like locations of the Alte Pinakothek with works not on display (Q30068277) (I just created the item). But that can also be done later of course. --Marsupium (talk) 12:34, 27 May 2017 (UTC)

BTW: How did you solve the page 200 threshold problem? And would you be OK if I'd reuse your code? You have'nt given a licence for https://github.com/multichill/toollabs/blob/master/bot/wikidata/bavarian_import.py if I did not miss it … (just the print ( u'Dude, write your own bot')) I'm not sure if I can find the time this summer anyway but if so, I'd do a try parsing and mapping the "Material / Technology / Carrier" field and adding that information here perhaps after further developing our technique/material ontology for the information from there and also from c:Template:Technique. --Marsupium (talk) 12:34, 27 May 2017 (UTC)

Creator again

Hi Marsupium, I worked on the creators. I would like to get the remaining paintings connected to a creator. https://tools.wmflabs.org/multichill/painters/index.php?collection=Q812285 is also updated so plenty of links to the Pinakothek. We still have the anonymous work challenge. I would propose:

1. Let's get all paintings linked up with a creator. This leaves us with some paintings to fix
2. Fix all the incorrect creators (see list on Wikidata:WikiProject sum of all paintings/Collection/Bavarian State Painting Collections to fix). This seconds step can probably be done with a bot after the first step is done

What do you think? Do you want to help? Multichill (talk) 17:39, 5 November 2017 (UTC)

@Multichill: Sure! I'm uncomfortable with those creators since about half a year :), always planning to write a commit for your bavarian_import.py to handle the anonymous cases, but failed and didn't try hard enough to even set up a proper test environment with the pywikibot module search paths … In the long run it might still be worth the work, I expect their (sometimes messy) website to improve in days to come. The two steps seem reasonable. I might get a bit more busy off-wiki now, but I will concentrate on getting down https://tools.wmflabs.org/multichill/painters/index.php?collection=Q812285 and https://tools.wmflabs.org/mix-n-match/#/list/467/unmatched when here! Thanks for bringing this up! --Marsupium (talk) 20:43, 6 November 2017 (UTC)

Great. I'll have a look at the bot at some point. I think the website changed and the problem might have disappeared with it. Multichill (talk) 22:19, 6 November 2017 (UTC)

@Multichill: I have yet reduced the backlog a little bit. This encoding error is though a task for a bot, perhaps you have an idea and can look into it to improve it for future imports! :) --Marsupium (talk) 22:13, 15 November 2017 (UTC)

Images and other fixes

sample upload

Hi, I figured out the source of the attribution stuff in the title: The API. In the API the attribution like workshop of is included in the title and not in the author. I switched both to parsing the page and I also added html escaping to get rid of the funky characters. You could just drop the current labels/descriptions on Wikidata:WikiProject sum of all paintings/Collection/Bavarian State Painting Collections to fix and have the bot re-import it now. Not sure what the best option is.

I'm also importing image suggestions. Prototype to upload these to Commons is at Commons:User:Multichill/Paintings to upload/Licensed-PD-Art. Multichill (talk) 19:05, 24 January 2018 (UTC)

Hi there! Wow, I didn't know there is an API, perhaps it's new!? Can you import the creators to our creator qualifiers? If so, we could indeed delete current titles and creators where the after fine-tuning the regex – I've improved it a while ago, I could do the rest. I'd prefer to have a clean import as a more sustainable way also for future changes of the data and non-painting artworks import some day …

Thanks a lot for the new image importing functionality, it is genious! Perhaps we should add section headings (in the SPARQL) for the file description pages? --Marsupium (talk) 20:28, 24 January 2018 (UTC)

The had an api when I did the first import. See for example https://www.sammlung.pinakothek.de/api/search?&page=1&filters={%22genre%22:%22malerei%22,%22publicDomain%22:true} . Default is xml, but it can also return json (check out my source). I can't really handle the different kind of attributions right now. You can see in this edit that it does end up in the right field now.

I'm not a big fan of the section headings. I often leave them out. Might add them again at some point. These pages are just prototypes. If all goes well I'll have a bot automatically upload all the easy cases on a regular basis so the humans can focus on the harder edge cases. Multichill (talk) 20:42, 24 January 2018 (UTC)

Okay, then I don't know if that should be done. What is bad, that atm we have wrong data, but it is easier to port (currently misplaced) creator QIDs than having to match them again from item descriptions, right? I'm not sure what should be done. --Marsupium (talk) 20:49, 24 January 2018 (UTC)

Creator issue has reached Commons

@Multichill: For example at File:Philips Wouwerman (zugeschrieben) - Vor dem Ausritt zur Falkenjagd - 1034 - Bavarian State Painting Collections.jpg the creator issue has now reached Commons, it should have {{Creator:Philips Wouwerman|attributed to}} with the attributed to as first parameter, but unfortunately hasn't … Perhaps Commons BStGS uploads should be stalled before this awful flaw is fixed. --Marsupium (talk) 22:36, 13 February 2018 (UTC)

Joseph Lowry

I've had a quick check on Joseph Wilson Lowry (Q6288013) - this is indeed Oxford Dictionary of National Biography ID (P1415):17102. The problem is that his entry is bundled together by the ODNB with the one for his father, Wilson Lowry (Q8023126) (17103). and their index pages are a bit confusing as they only list the citation for "main" article subjects, not anyone else dealt with in the same article but a different ID. Definitely the right ID, though. Andrew Gray (talk) 21:51, 6 September 2017 (UTC)

Oh sorry! I've yet encountered these confusing IDs once, but for forgot about it. Thank you for your explanation! --Marsupium (talk) 21:56, 6 September 2017 (UTC)

Thanks

Thanks for having fixed my error on PA_005 (Q28470238), understood. GAllegre (talk) 16:57, 19 September 2017 (UTC)

Thank you for your answer. I think it would be good to have that functionality to use arbitrary images that's why I once proposed to create a second property to be used as a qualifier on image (P18) statements. We could propose it another time, however, even the existing property isn't that popular yet, perhaps because there aren't any simple tools to get the numbers. So it might be a thing for the future. --Marsupium (talk) 17:03, 19 September 2017 (UTC)

dedicated to (P825)

Hi,

I've just removed your example from this - as I understand it, dedicated to (P825) is for the dedication on works - "I dedicate this novel to XYZ". Buildings in honour of someone, or religious dedications to saints/deities etc, usually use commemorates (P547) (& named after (P138) if they're named for them). Andrew Gray (talk) 11:24, 27 September 2017 (UTC)

Hi Andrew, actually this was a kinda trial balloon testing if anybody opposes. :-) The label in English at least seems to fit, cf. w:Dedication. And church building (Q16970) is a work (Q386724), so is St. Peter's Basilica (Q12512) (otherwise it would raise a constraint violation, but it doesn't), I've looked it up this morning (UTC), but haven't the chain at hand now, you can find it.

I don't think commemorates (P547) or named after (P138) are quite fitting here. Commemoration isn't really the point neither is the naming. I was insecure about using dedicated to (P825), too. What let me add it were the types statistics indicating that out of the c. 7,658 uses many more than 6000 use the property in that way like on St. Peter's Basilica (Q12512). I haven't queried the statistics for commemorates (P547) and named after (P138) though, commemorates (P547) at least won't show this result certainly. I propose to move this small discussion to Property talk:P825 to make it visible for further discussion. Regards, --Marsupium (talk) 16:35, 27 September 2017 (UTC) 16:42, 27 September 2017 (UTC)

Good idea - I'll move this across now and reply there :-) Andrew Gray (talk) 11:31, 28 September 2017 (UTC)

Attribution qualifiers

Hello Marsupium. I saw your recents change but I am, like some others, really not conviced by the approach about attribution qualifiers. There is a discussion about that. The AAT approach, the same as the one that I encountered in museums, is imho preferable as explained for edition, referencing and reuse. Best regards --Shonagon (talk) 19:30, 2 October 2017 (UTC)

Hello Shonagon, I know the discussion. However, I think it is better to wait for its outcome before possibly starting with a new system and sourcing circumstances (P1480) won't fit even then probably. But I will stop changing stuff in this area before we have an consensus in that discussion. Thank you for reminding me! Best, --Marsupium (talk) 08:18, 3 October 2017 (UTC)

Help with Renaud de Montauban/Reinoldus

Perhaps you can advise me on Renaud de Montauban (Q727316). Do you think that we should have one Wikidata item for both the saint and the legendary knight, or should we have two items and link them with "said to be the same as"?

Based on the EN Wiki article, I would be inclined to split, but I'd like someone from DE Wiki to assess the problem from that perspective. - PKM (talk) 02:19, 7 November 2017 (UTC)

Hm, I don't know. What is with Reinold (Q7310416)? --Marsupium (talk) 16:56, 7 November 2017 (UTC)

WikiProject Coins

Good evening1 Can you add yourself into the Project Breg Pmt (talk) 20:49, 11 November 2017 (UTC)

Hi, I'm not so into coins actually, I only started to follow the project as someone interested in visual arts in general, but unfortunately I've never really worked with coins apart from those used in coin object (Q18536132) … I better concentrate on what I know more about I think. Thanks for your invitation tho! --Marsupium (talk) 20:55, 11 November 2017 (UTC) Ok but thanks anyway :) Breg Pmt (talk) 21:10, 11 November 2017 (UTC)

wondering... about place name sign (P1766)

Hello,

You reverted my edit.

AFAIK, this property has been used, consistently, on human (Q5) items, to link to places named after them, so I do not really understand your revert...

If this should not be the case,maybe it should be explicitly stated… --Hsarrazin (talk) 12:44, 12 November 2017 (UTC)

@Hsarrazin: It raised a constraint because of place name sign (P1766)property constraint (P2302)subject type constraint (Q21503250)class (P2308)geographic location (Q2221906) and appeares here. There seems to be a discussion start if to expand the domain of the property in the direction you've used it. I just stumbled upon the constraint violation, but your edit is in respectable society. I do not really care. Feel free to re-revert me! --Marsupium (talk) 12:52, 12 November 2017 (UTC)

thanks for the pointing on the discussion. It never occurred to me that there could be a problem with using this property this way. There is a real need to be able to have those signs linked to the person they have been named after. --Hsarrazin (talk) 13:00, 12 November 2017 (UTC)

Ok, I'd personally think it would be better to (create an item liked trough named after (P138) and) use place name sign (P1766) on the item for the geographic location. I don't really have a problem with using it the way you did though. But it should be documented properly, especially through the constraints. --Marsupium (talk) 13:06, 12 November 2017 (UTC)

Berlinische Galerie

Hi Marsupium, I imported the paintings of the Berlinische Galerie, see Wikidata:WikiProject sum of all paintings/Collection/Berlinische Galerie. Quite a few of these (German) painters don't exist on Wikidata yet, see https://tools.wmflabs.org/multichill/painters/index.php?collection=Q700222 . Want to help create them? The German Digital Library site contains GND/VIAF info for most people. So for example Arnulf Letto (Q48444177) has the work Q47522380 which contains a reference to https://www.deutsche-digitale-bibliothek.de/item/SVWUKZXZYBC5YM7FTPCZNWTFXGPC6OAS -> https://www.deutsche-digitale-bibliothek.de/entity/124637272 so these id's can be added too. Want to help? Bot will pick up some fields like date of birth and gender if it has an RKDartists link and will update the paintings with the creator. Multichill (talk) 16:23, 11 February 2018 (UTC)

Hi Multichill, thank you for the import! I've added some few items for Berlinische Galerie creators. A problem with perhaps half of the missing creators is that the names from the import include dates of birth and/or death which prevents the bot from merging them. --Marsupium (talk) 19:20, 12 February 2018 (UTC)

Yeah, noticed that too. I wanted to wait a bit for the real missing creators to be created and than fix them. It seem have happened for artworks where the api didn't report a person item. So please, do cherry picking and worry about the difficult ones later :-) Multichill (talk) 21:04, 12 February 2018 (UTC)

That's what I'm doing! :-) Have a nice evening! --Marsupium (talk) 21:05, 12 February 2018 (UTC)

Merci beaucoup

Bonsoir, c'est moi qui vous remercie pour ces conseils et la correction de mes erreurs de débutant. Je vais essayer d'en faire le moins possible. Si ce n'est pas abuser de votre temps, pourriez-vous m'apprendre à créer des déclarations concernant la qualification des artisans d'art français avant la Révolution. En effet ils étaient:

1. 1. ou ouvrier libre
   2. ou reçu maître et
2. cela (localisation)
   1. à Paris
   2. ou plus rarement dans une autre ville
3. à telle date et
4. selon 3 circonstances :
   1. après chef d’œuvre
   2. en tant que fils de maître
   3. et plus rarement, notamment pour les non français, par décision royale.

Merci d'avance. Finoskov (talk) 21:15, 28 May 2018 (UTC)

Salut Finoskov! (J'ai numéroté les détails dans ta contribution (j'espère que c'est ok!) et modifié ma réponse:) Est-ce que tu as un exemple pour cela? Peut-être déjà avec des sources? J'ai vu que tu as utilisé occupation (P106), par ex. ici. Je pense que position held (P39) serait meilleur que occupation (P106). master craftsman (Q1284709) n'est pas une occupation ou profession, mais une fonction ou un titre. (Et nous devrions changer le instance of (P31) de master craftsman (Q1284709) en position (Q4164871).)

Alors voici ma proposition:

1. Deux variantes:
   • ou position held (P39) (mon option préférée): John Doeposition held (P39) ouvrier libre / master craftsman (Q1284709)
   • ou significant event (P793): John Doesignificant event (P793)nomination ouvrier libre / nomination maître de guilde
2. location (P276) est bon.
3. point in time (P585) comme qualificatif n'est pas correct pour position held (P39) (ou occupation (P106)) ici comme tu l'utilises, il doit être start time (P580). (Pour significant event (P793) le qualificatif point in time (P585) serait bien comme ça toutefois.)
4. has cause (P828) (Vois aussi Help:Modeling causes.): John Doeposition held (P39)master craftsman (Q1284709)has cause (P828) après chef d’œuvre / fils de maître / décision royale
   1. for work (P1686) pourrait utile ici, on devrait changer la portée ou l'application recommandée pour l'usage un peu possiblement.

C'est pas facile, tout ça … :) J'ai aussi demandé aux autres ce qu'ils en pensaient. S'il te plaît laisse-moi savoir si tu es d'accord ou non! Merci d'avance! Cdlt, --Marsupium (talk) 08:48, 1 June 2018 (UTC), 13:27, 6 June 2018 (UTC)

P.-S.: Au fait, tu fais mauvais usage de field of this occupation (P425), vois notamment le contrainte de type et le contrainte de type pour la valeur! --Marsupium (talk) 14:09, 6 June 2018 (UTC)

Bonsoir, n'étant pas réellement anglophone, j'ai du mal à apprécier la différence exacte, pour wikidata, entre le mot anglais occupation (qui est quand même mis en relation dans la description wikidata avec statut, métier, profession, ce dernier étant d'ailleurs la traduction en français du dictionnaire Larousse pour le mot anglais occupation) et fonction, mot français, pouvant désigner, entre autres, le métier, la profession. C'est ce qu'on appelle en linguistique un faux ami. Personnellement je n'ai pas de préférence (il suffit de prendre une décision), même si en français le terme générique le meilleur pour métier, profession est Fonction, ce qui obligerait à lui attribuer certaines propriétés liées actuellement à Occupation. Je suis donc de l'avis de Marsupium mais pas pour la même raison puisque la Maîtrise est bien une qualification professionnelle artisanale. En effet, avant la Révolution, la maîtrise était indispensable pour exercer en nom propre, (c'est à dire de façon indépendante, pour son propre compte), une profession artisanale, ou un métier d'art. C'était une sorte d'examen professionnel qualifiant comme on dit de nos jours, alors que les titres comme : horloger du roi, de la reine et autres , étaient eux des titres et des fonctions honorifiques.

Pour éviter cette ambiguïté linguistique entre Fonction et Occupation la solution suggérée par Marsupium d'utiliser significant event (P793) n'est-elle pas plus simple à mettre en œuvre?

D'autre part le statut d'ouvrier libre étant assez rare et uniquement, semble-t-il, à Paris, on pourrait l'oublier afin de ne pas compliquer les choses.

Cordialement et encore merci à Marsupium. Finoskov (talk) 20:28, 6 June 2018 (UTC)

!

IT MAY SURPRISE YOU BUT i DO CHECK FOR DUPLICATES BEFORE CREATING A NEW ITEM. What I had done there was create a list of all people called John Knox that were in Wikidata, then go through the list at Mix 'n Match to see whether they were on the list, or whether there were multiple mentions of people not on the list. This one was not on the list, so I created it. To go to such depths and THEN be told that I am doing something wrong because I did NOT do what I DID do VERY EXTENSIVELY. To be honest, I should not have gotten that angry, but it really isn't surprising either. - Andre Engels (talk) 08:15, 14 July 2018 (UTC)

Apologies accepted, and again sorry for overreacting in this way. - Andre Engels (talk) 08:36, 14 July 2018 (UTC)

(EC) @Andre Engels: I'd like to apologize from my side, too! I didn't know about your extensive work on it! I was a bit in a rant after spending once more a while on merging a bunch of in my eye uber-obvious dupes. I remember I looked who added "Jack Knox" as an alias and stupidly probably was to hasty and thought it was you yourself. Sure you don't have to look for John Knox when you create a page about Jack Knox! You may en:WP:SHOUT, I'd appreciate if you'd remove the en:WP:PA above tho. Once again sorry for my carelessness and anger! --Marsupium (talk) 08:43, 14 July 2018 (UTC)

Thanks, and you're right, this kind of words is not only not conductive to discussion, it's also inappropriate. I have edited my two messages into one, removing the irrelevant parts. - Andre Engels (talk) 08:54, 14 July 2018 (UTC)

Thank you and happy editing! --Marsupium (talk) 09:06, 14 July 2018 (UTC)

Reverting a three year old entry is neither nice nor needful

simply delete it, so you do not disturb other authors. Or do you know what status quo a property has had in 2015? --Plagiat (talk) 16:24, 31 August 2018 (UTC)

@Plagiat: I'm sorry! No, I don't. I don't revert to tell some edit of 2015 has been wrong then. I personally like to get notifications when something I edited is deleted. I'm learning it's different for others and I see it was an half-automatic edit of yours … So sorry for disturbing! --Marsupium (talk) 18:07, 31 August 2018 (UTC)

tombstone (Q203443), crucifix (Q20460) and sculpture (Q860861)

Hello User:Marsupium. Your recent ontologic changes on tombstone (Q203443) and crucifix (Q20460) are welcome. In consequence, many headstones and crosses are not considered as subclasses of sculpture (Q860861) by default. That was indeed an issue, resolved now. A little new issue is that some of them are scluptures too. So, for information, I added instance of (P31):sculpture (Q860861) for those which are obvisouly scluptures too. Best regards --Shonagon (talk) 12:01, 2 September 2018 (UTC)

Salut Shonagon! Thanks a lot for your feedback and for pointing out the issue with the “sculpture crucifixes” and for your work on fixing it! Merci beaucoup, cdlt, --Marsupium (talk) 12:40, 2 September 2018 (UTC)

Saddle bags

The second is called Heyba, it is similar to Khurjun (which is first) but it is different thing. In English both of them are saddle bags. --Interfase (talk) 18:15, 8 September 2018 (UTC)

I'd like to keep the discussion at one place, thus I've copied your reply to Topic:Ukd5auglbtgrs20a and replied there. --Marsupium (talk) 18:45, 8 September 2018 (UTC)

botanists named hooker

There are two or three Hookers who have written notable scientific papers and have recognized author abbreviations. This one has "Hook. fil." (wikipedia says Hook. f. which is more recent) auth abbrev. when citing his works. en:Joseph Dalton Hooker, en:William Jackson Hooker. So, it is not a punctuation so much as it is an abbreviation, of what, I do not know. I put it there to help when finding it via search here. My grammar is off while juggling the different areas. And I seem to have just stopped sentences when I was unable to figure out the right word. Thank you for suffering through this! --RaboKarbakian (talk) 23:47, 10 October 2018 (UTC)

Hi RaboKarbakian, you're talking about this edit of mine, right? I'm sorry, I had already realized the meaning of the period and reverted my edit. Or have I removed any other periods? :o Feel always free to revert me! Cheers and happy editing, --Marsupium (talk) 00:00, 11 October 2018 (UTC)

Yes, that edit. I was glad to know there was a rule for it, truthfully.

I have looked up the word lithography. I am proud to be the grandchild of a photolithographer! I am wondering if I should call them lithographs instead of simply a print. Do you know if there is any one currently making lithographs in that old way. While working on them, the last thing that was in my guesses about how they printed them was that they used a wet stone....--RaboKarbakian (talk) 00:05, 11 October 2018 (UTC)

Lithography is a still practiced artistic graphic technique, Christo and Jeanne-Claude (Q158477) for example used/uses it. Its use for everyday prints is probably not very common, but I dunno. You can probably also use "lithograph" even without a source for that, because quite sure they are. --Marsupium (talk) 00:31, 11 October 2018 (UTC)

Is Japanese painter project for LACMA closed?

If you are still interested on Japanese painters for a LACMA collection, kindly ping me please. Cheers, --Omotecho (talk) 06:10, 31 October 2018 (UTC)

Thanks for replying at Wikidata:井戸端#Identification or creation of items for Japanese painters, I've in turn replied there. --Marsupium (talk) 17:03, 31 October 2018 (UTC)

Getty-Wikidata import

Thanks Marsupium for the friendly welcome on the Project chat page. My team and I would like to make sure we coordinate with your GVP efforts so that we may avoid any pitfalls and conflicts. Our plan at the moment is to submit our local authority data (architects and buildings) to Getty's ULAN and CONA programs, and then, via OpenRefine, update any matching Wikidata records with Getty IDs if needed, and create new Wikidata records with Getty IDs if no matches are found via the reconciliation process. We are working with a large enough dataset (~10000 records) that we will be using a number of quality assurance tricks to prevent duplicates, especially those described at Wikidata:Tools/OpenRefine/Editing/Quality_assurance. Let me know if you have any other comments and suggestions about our effort. Im happy to keep you informed about our progress on your talk page here, or elsewhere.--Dzahsh (talk) 17:30, 12 February 2019 (UTC)

@Dzahsh: Hi, that sounds great! It's the best result when your data gets both into GVP and Wikidata. And submitting first to ULAN and CONA and from there to Wikidata sounds reasonable. And architects here need definitely more work compared to e.g. painters. :-)

I'm also working on deduplication of visual artist (Q3391743)s based on labels, occupation (P106) and date of birth (P569)/date of death (P570), so perhaps I can chip in with that for reconciliation.

Unfortunately, I'm not too familiar with OpenRefine, but I've already planned to familiarize myself with it more. Maybe that will also work better for GVP → WD than my current workflow that's based on sequential sets of SPARQL queries and corresponding QuickStatements batches. It would be great to find instruments for the GVP → WD step that can be applied maybe to ULAN and CONA in general, let's see! CONA import might become more difficult than ULAN, since Getty doesn't provide it at their SPARQL endpoint yet.

Ah, I'm still wondering what kind of data you have, are the buildings the bigger part of the 10,000 records? And which temporal and geographical range do they cover? Probably there aren't any samples public yet, right?

I'd be pleased to hear from you about the progress on the project! (On this page works well for me.) Best, --Marsupium (talk) 01:14, 14 February 2019 (UTC)

@Marsupium: Our data is about half buildings and half architects and architecture firms. We are still preparing it so there's nothing public yet. Would you recommend that I set up a project page somewhere describing my progress? I feel like it might be useful for future Getty-Wikidata projects --Dzahsh (talk) 19:07, 20 February 2019 (UTC)

@Dzahsh: Thank you! Sure, create a project page! I think in most cases it's the more public documentation the better. --Marsupium (talk) 15:01, 25 February 2019 (UTC)

Jena Codex

Ad Your revert https://www.wikidata.org/w/index.php?title=Q12024512&diff=868190492&oldid=804103921: I realized, that I am not able to exactly describe this problem within Wikidata item, so I write directly to You: Jena Codex is convolut and it contains Incunabula (I think, that we call it Adligat (Q16526512)) as well as part with hand writing. Incunabula contains Addendum to Voragine Legenda Aurea with "czech saints" ("Jena Addendum"), cf. first page [4] (clearly printing, not manuscript); in czech bibliography "Knihopis" under catalogue number INC020 [5] or in ISTC (but it is more general and less exact) ij00138000 [6]. --marv1N (talk) 16:22, 27 February 2019 (UTC)

Thanks a lot for your message and for pointing this out! And sorry for the revert for wrong reasons. I still don't think Jena Codex (Q12024512) should have instance of (P31)incunable (Q216665), incunable (Q216665) should rather be used on instances of version, edition or translation (Q3331189), I will model it later as I think it should be and let you know and you can tell me if you agree. Thanks again! --Marsupium (talk) 11:42, 28 February 2019 (UTC)

Yes, I didn't reverted back, because it should be expressed better (somehow :-)). instance of (P31)miscellaneous manuscript (Q2217259) is more or less sufficient, but combination of handwriten part and printed part is "little" complication (on the other hand Jena Codex isn't the only miscellaneous manuscript (Q2217259) containing this combination - so, Wikidata are suposed to have some functional model for this cases - thanks for modeling it...). --marv1N (talk) 15:13, 28 February 2019 (UTC)

@Marv1N: Now I've used exemplar of (P1574) on Jena Codex (Q12024512), it still isn't a good description, also because I haven't completely understood what all is in that codex, but in principle this is a clean way to connect the edition and the specific exemplar. Do we have an item for the work the manuscript is an exemplar of? --Marsupium (talk) 19:56, 7 March 2019 (UTC)

I hope, it is sufficient (at this moment, there could be discussion, if codex (Q213924) is exemplar of (P1574) of incunable (Q216665)). When it is convenient, I use P1574 as well (but, as You said, we need item for each work). After some searching, I finaly discovered more or less complete listing of works in Jena Codex: http://www.encyklopedieknihy.cz/index.php/Jensk%C3%BD_kodex (our incunabula is "39r-54r: inkunábulový dodatek k Pasionálu (1495)"). I don't see any of these works on Wikidata at the moment. --marv1N (talk) 10:32, 8 March 2019 (UTC)

Ok, thank you, creating the items isn't so easy for me because of the Czech descriptions. I've added the link to Jena Codex (Q12024512) not to lose it though. Hopefully, the item will improve over time. --Marsupium (talk) 18:03, 8 March 2019 (UTC)

Name change

Hello.

Astana International Airport has been officially renamed to en:Nursultan Nazarbayev International Airport.

Please see inside the official website of this airport: http://nn-airport.kz

Can you rename the German article?

Yours sincerely, Noraton (talk) 15:51, 11 May 2019 (UTC)

@Noraton: de:Flughafen Nursultan has been moved to that lemma in last month. No idea what I should do here. --Marsupium (talk) 20:35, 11 May 2019 (UTC)

Hello.

The airport needs to be renamed as de:Internationaler Flughafen Nursultan Nasarbajew because the airport officially bears the first name and the surname of the former president.

There is an example in German Wikipedia from Swaziland where the airport is named after the current King, de:Internationaler Flughafen König Mswati III..

German Wikipedia allows this.

Yours sincerely, Noraton (talk) 18:54, 12 May 2019 (UTC)

@Noraton: I don't have much experience with dewiki's rules on airport article titles, but following de:Portal:Luftfahrt/Namenskonventionen#Artikeltitel für Flughäfen und andere Flugplätze and http://www.ans.kz/AIP/eAIP/2019-05-23-AIRAC/html/index-en-GB.html, AD 2.2, 6 I've moved the page to de:Internationaler Flughafen Nursultan Nasarbajew now. --Marsupium (talk) 19:31, 12 May 2019 (UTC)

Thank you very much! Noraton (talk) 19:47, 12 May 2019 (UTC)

grammatical gender: masculine on German nouns

Hi Marsupium! I hope this doesn’t sound rude, but… what’s the deal with these grammatical gender (P5185)masculine (Q499327) statements you seem to have added recently? Earlier today I already reverted you on Anarchäologie (L1839), but I assumed that might be a outlier (I’ll admit it’s a weird word) – but now I just opened some random lexeme, Ballspende (L40632), and found the same thing here: I’m not sure what a Ballspende is supposed to be, but I’d consider it to be feminine. Is this a larger problem? --Lucas Werkmeister (talk) 17:35, 6 July 2019 (UTC)

Hi Lucas Werkmeister! Thanks a lot for the corrections and the revert of the first case! Of course you're right. I cannot really account for how that happened. Either I wasn't of sound mind someway or it was a technical problem sometimes happening to me where when you click on a suggestion in the dropdown menu for entities the menu disappears and not the clicked entity is inserted but an entity selected (e.g. with using the arrows). I'm aware of this phenomenon, but maybe in those cases it escaped my attention. I've filed phab:T227387 for that.

I've checked all the 17 additions of grammatical gender (P5185) I did and the others seem to be ok. Sorry for the disruption and thanks again for noticing and pointing it out to me! --Marsupium (talk) 19:45, 6 July 2019 (UTC)

@Marsupium: Thanks! Just wanted to make sure it wasn’t some automated thing affecting hundreds of lexemes – sounds like I just got very unlucky (or lucky?) to stumble across both of these cases within a day! --Lucas Werkmeister (talk) 22:07, 6 July 2019 (UTC)

commons:Category:Guruy executes Siyavush

Hi, I saw your revert of Pi bot. I think what went wrong was that it added the Commons category to a painting of an event, while the category is about the event. Can you confirm? If possible, could you create a Wikidata item for the event to describe the Commons category? Thanks. Mike Peel (talk) 17:11, 27 May 2020 (UTC)

Thanks for noticing! Confirmed and item for the scene Guruy executes Siyavush (Q95841503) created. --Marsupium (talk) 11:33, 30 May 2020 (UTC)

A long time ago in a property far, far away...

This looks like a tricky case: On 2018-11-06 you (correctly) reverted two edits from 2018-01-08 making instance or subclass of (Q30208840) a subclass of (P279) instance of (Q21503252) as well as subclass of (Q21514624).

Then on 2019-05-02 a non-logged-in user from a Canadian IPv4 address did a restore of the same item to what it looked like after one of three edits made on 2018-08-28, thereby effectively reverting all 15 edits made during the previous eight months, resulting in the loss of labels, descriptions or aliases in eight languages (be-tarask, es, fi, he, id, nn, sl, sv). And the edits you had reverted were back from the dead.

I find this interesting for multiple reasons:

1. The revision restored was an intermediate one that hadn't lasted for even a minute; it was the first of three edits made by one user, the addition of a label in Spanish. The Spanish description and alias added by the same user were lost in the restore. Therefore I find it highly unlikely that this particular user had anything to do with the restore. It looks more like a random pick of any entry found in the revision history of this item.
2. The restore was tagged "Mobile edit", "Mobile web edit", which seems consistent with the logging of an essentially anonymous IP address. I'm unfamiliar with the criteria for these two tags being added; do they depend on the actual interface software used (and thus relate to its graphical expressiveness) or is it an interpretation of the device characteristics as they are reported by the HTTP protocol?
3. I view the item affected as kind of semi-important due to its role in the interpretation of property type constraints. Its exact properties maybe aren't critical to the daily operation of Wikidata, and it's not semi-protected either as far as I can see. Maybe it should be?
4. This clearly inappropriate restore has apparently gone unnoticed for 15 months (I'll explain further down what made me find it). Even mere claim changes are separately tagged as kind of suspect ("New user changing statement") if the user isn't auto-confirmed yet; why not restores? Or are anonymous IP addresses exempt from the "new user" semi-quarantine? If so, why? I'll let someone who is more familiar than me with Mediawiki determine whether this is a design bug or not, and find out where to report it; my analysis here is free for anyone to use.

In summary, I find it odd that an anonymous user effectively can erase months of work on an item with a restore operation from a mobile web interface (if that is indeed what it is) without an easily discernible signal being sent anywhere (it doesn't have to make Obi-Wan Kenobi sit down as if the disintegration of an inhabited planet has caused a sudden fluctuation in the Force, but maybe it can make R2-D2 connect to the waste compressor to find out what was just flushed down the kitchen sink), and maybe even by mistake. In case there was a confirmation step, the anonymous user may have misunderstood it, perhaps thinking it was a local page refresh operation. Or just tapped the wrong spot on his touchscreen without even noticing (I do that sometimes, it's utterly annoying and I have disabled the touch pad on my regular work laptop, using a wireless mouse instead).

I have no objection to non-logged-in users adding or changing labels, descriptions or aliases, maybe even deleting them (an existing alias could be totally inappropriate, and we shouldn't discourage anyone from removing foul language from item descriptions). I'm a bit more skeptical about doing non-logged-in reverts (undo) of the singular edits of others, though we should probably allow it (I recall I actually used that possibility once to revert vandalism in a Wikipedia article without logging in; maybe I was concerned the vandal might otherwise find out who did the revert and come after my Wikipedia handle for revenge).

But restores? Before doing my first revert, I recall I had trouble understanding the difference between undo and restore; they seemed synonymous to me. I didn't immediately find a Help page explaining the difference, but I picked undo as it sounded less dramatic, and it turned out to be the proper choice. I have never done a restore myself even when logged in (as far as I can recall, but if I did, it must have been to a very recent revision, effectively the same as an undo of the few most recent edits, certainly not as much as 15 edits in one step).

In my opinion, restores should not be possible for non-logged-in or not yet auto-confirmed users, as they can effectively be the same as erasing the entire item. Is erasing an item in Wikidata even allowed for regular logged-in users? Aren't we supposed to turn old items into redirects to wherever their contents have gone in case of a merge?

Now, I don't know whether you are involved with these issues, but I begin here as you have first-hand knowledge of some of the edits that were affected by this incident. I considered pinging either SR5 or Jura who seem to have been more involved with item creation or the real property (rather than language) edits, but decided to give you the first opportunity, and let you decide who else to notify (if anyone, but please let me know if you don't think this is worth spending any effort on and why, so I can decide whether I agree or not).

As to the 15 lost edits, the inappropriate restore has since been followed by 29 edits involving even more languages (including some of the lost ones) plus a couple of legit Wikidata property (P1687) claims made by SR5, so I don't think there is much point in undoing the restore itself (if that is even possible now), but it's probably easier to go through the lost languages manually and repair those that haven't been fixed yet one by one (I could perhaps do that if you don't have the time). And you get the opportunity to revert the very same original mistaken edits a second time!

As to how I found it: I noticed an unusual alert on the field of work (P101) item subject type constraint (Q21503250) indicating relation (P2309) had more than one value (instance of (Q21503252) and subclass of (Q21514624)). Thus I learned about the purpose of instance or subclass of (Q30208840), and assumed it was a parent class of the two now conflicting relations. Wrong, Wikidata told me; it was a subclass of both! What..? Thinking I might be able to give someone a free lesson in class logic, I looked into the revision history and saw you had removed those claims long ago. So, why were they still there, as nobody else seemed to have added them? Ah, a restore, how quaint...

Now, while you and others may ponder the implications of my analysis above (if there is anything of interest in it), I think nobody will be upset if I repair the broken P101 type constraint (after careful consideration of its revision history, talk page and other potential sources of relevant information on the subject), after which I can perhaps find out whether field of work (P101) is really meant to be used also with items that are instances of mathematical concept (Q24034552), such as intersection (Q17141489) which, surprisingly to me before this investigation, isn't flagged (and before you ask, it's not due to it being an operation (Q3884033), as I added that one myself earlier today). Personally, I would prefer studied in (P2579) for this property value, and will probably make it so after I have had something to eat. --SM5POR (talk) 20:48, 1 August 2020 (UTC)

@Marsupium: I'm doubtful as to whether you have seen my report above; doing an explicit ping just in case... --SM5POR (talk) 19:57, 3 August 2020 (UTC)

A New Hope

@SM5POR: Sorry for the late reply! Was lacking the leisure to read through your message until now, sorry. I've fixed as many as possible of the edits destroyed by the restore. I think I can't contribute much to the more general questions on who should be able to perform what action. Cheers, --Marsupium (talk) 16:12, 25 August 2020 (UTC)

@Marsupium: Thank you! No worries, if it has so far waited 15 months for anyone to pay attention, it can wait a few more weeks (and I have been busy too). I'll see if I can bring it to someone else's attention then. But since my report is kind of bulky, would you allow us to continue discussing the issue here on your talk page, at least until someone authoritatively tells me where to file it? I took the liberty of adding a sub-section heading above to simplify further comment edits below; I hope you are ok with that too. It's the anonymous restore issue that is my primary concern, not the particular item where I found it. --SM5POR (talk) 12:04, 27 August 2020 (UTC)

Call for participation in a task-based online experiment

Dear Marsupium,

I hope you are doing good,

I am Kholoud, a researcher at King's College London, and I work on a project as part of my PhD research, in which I have developed a personalised recommender system that suggests Wikidata items for the editors based on their past edits. I am collaborating on this project with Elena Simperl and Miaojing Shi.

I am inviting you to a task-based study that will ask you to provide your judgments about the relevance of the items suggested by our system based on your previous edits. Participation is completely voluntary, and your cooperation will enable us to evaluate the accuracy of the recommender system in suggesting relevant items to you. We will analyse the results anonymised, and they will be published to a research venue.

The study will start in late January 2022 or early February 2022, and it should take no more than 30 minutes.

If you agree to participate in this study, please either contact me at kholoud.alghamdi@kcl.ac.uk or use this form https://docs.google.com/forms/d/e/1FAIpQLSees9WzFXR0Vl3mHLkZCaByeFHRrBy51kBca53euq9nt3XWog/viewform?usp=sf_link I will contact you with the link to start the study.

For more information about the study, please read this post: https://www.wikidata.org/wiki/User:Kholoudsaa In case you have further questions or require more information, don't hesitate to contact me through my mentioned email.

Thank you for considering taking part in this research.

Regards

Kholoudsaa (talk) 14:08, 8 January 2022 (UTC)

Atlas ID (P1212) No formatter URL in English

Hello Marsupium On Atlas ID (P1212) I removed the formatter URL for English because there was no English version for all notices. The ID matched all notices in French only which had sometimes English versions, or Spanish in some rare cases, with the same ID. The database is closed now but the information is still relevant for references or web archives. Best regards --Shonagon (talk) 21:45, 16 January 2022 (UTC)

Shonagon, thanks for the message! The URLs now redirect to the new ARK IDs and the redirects (I've tested) respect the chosen language. So could we now keep the English formatter URL or are there still URLs for English that don't work and we should remove it again? Best and thanks, --Marsupium (talk) 23:02, 16 January 2022 (UTC)

Hello Marsupium. Yes with the redirection from Atlas to new ark ID takes into account the language parameter. And indeed in that way, the formatter URL for English is usefull. The point is that that these references to English Atlas notices could have never existed. That would be a problem if someone make a search in web archives for one of this no preexisting notices. Therefore, yes, for redirections, if there is one in French there will the same for English. --Shonagon (talk) 12:00, 22 January 2022 (UTC)

AKL-Import

Hallo Marsupium, du hast Paulus Juvenel the Younger (Q116764874) mit der AKL Online Künstler ID 20009578 importiert. Wenn du sie oben ins Suchfeld eingibst findest du Paulus Juvenel the Younger (Q55853106). Auch GND 137672381 "added with more Identifiers" by User:Hannolans ist bereits in Wikidata erfasst. --Kolja21 (talk) 20:48, 24 February 2023 (UTC)

Vielen Dank für den Hinweis! I could imagine that I didn't find Q55853106 because I only searched for "Paul Juvenell" without "-us" and with double "ll". Hannolans has already merged the items. Thanks for doing so! --Marsupium (talk) 22:26, 25 February 2023 (UTC)

Kunstdenkmäler-Import

Hallo Marsupium,

hast du das Handbuch der deutschen Kunstdenkmäler, Teil: Nordrhein-Westfalen / 2., Westfalen (Q117286835) digital vorliegen und importierst du alles daraus?-- U. M. Owen (talk) 12:11, 15 September 2023 (UTC)

Hallo U. M. Owen,

leider nein. Für die Objekte, die ich angelegt habe, habe ich die Daten per Hand aus der Druckausgabe übernommen. Es gibt das Projekt Digitaler Dehio (bisher offenbar ohne Properties für IDs in Wikidata), in dem die Region aber, soweit ich es überblicke, (noch) nicht abgedeckt ist. Hättest Du für einen bestimmten Zweck Bedarf an (Auszügen aus) dem Band? Beste Grüße, --Marsupium (talk) 20:15, 15 September 2023 (UTC)

provenance loves wiki – Berlin

Hallo Marsupium, vielleicht kannst Du auch kommen: https://de.wikipedia.org/wiki/Wikipedia:Arbeitsgemeinschaft_Kunstwissenschaften_%2B_Wikipedia/Provenance_loves_Wiki/English Das wäre schön. Viele, Grüße Pippich Pippich (talk) 11:04, 6 October 2023 (UTC)