Wikidata talk:WikiProject Visual arts/Item structure/Art movements
I think the basic concept of the project is great in all it's simplicity. The complexity will follow with the content!
I have some suggestions for the ListeriaBot. They can be easier or more difficult to implement:
- Have multilingual text in the table. Should there be a variant of the {{Q}}-template without the appended Q-ID to make it shorter?
- Could the table be made editable with menus, translate-shortlinks etc? I don't know the status of this in Wikidata.
- Commons-category could be a link.
In general, when the page settles, it would be great to translate it. In my mind an important effort from local art buffs would be to provide the terminology used in the local language (not that they would not understand English).
Looks good edit
Thanks for tackling this. I've been frustrated because I didn't really know where to start. - PKM (talk) 01:23, 27 July 2015 (UTC)
How about adding LIDO edit
great initiative, thanks. Checking through, I wonder if it would be of use to integrate LIDO to the table as well. What would it mean workwise? Where to start? I could not say, but LIDO is in standard use for metadata within European museums.--Barbara Fischer (WMDE) (talk) 10:33, 28 July 2015 (UTC)
- @Barbara Fischer (WMDE): I know LIDO (started the en:LIDO page) and would be interested in any LIDO work. But can't figure out how is it related to Art movements, can you explain? --Vladimir Alexiev (talk) 17:07, 31 July 2015 (UTC)
Ethnic groups edit
As part of Europeana Food and Drink we're working with some ethnographic objects: thus cultures (eg Ainu), and some more modern objects: thus periods (eg Victorian). As part of that work we're looking to coreference the BM ethName thesaurus (also used by the Horniman museum), AAT Periods and Styles and DBpedia (thus Wikidata). Thanks for the list of Periods below, that'll help in our efforts! Of course, we'll publish it on WD as AAT ids, or in Mix-n-Match.
Re the difference ethnic group/culture/people/style/period/movement: I think Getty slapped them all together in one facet for a good reason. I'm no art historian, but it seems to me the distinctions between them are often blurry. So I hope we're not painting ourselves into a corner by instituting all these different properties. Cheers! --Vladimir Alexiev (talk) 17:13, 31 July 2015 (UTC)
- Hi @Vladimir Alexiev:, you are totally right in pointing this out. I have taken some time to investigate this more closely. Indeed, the Art and Architecture Thesaurus of the Getty puts all this in one 'tree' and I agree with that decision. So I have slowly started editing non-Western art movements, such as Igbo art (Q5991589), to reflect this. Unfortunately Wikipedia is still very bad at covering non-Western art, so we still have a long way to go. We need to start somewhere though. Spinster (talk) 19:28, 18 September 2015 (UTC)
Matching Periods and Styles edit
Notified participants of WikiProject Visual arts
Hi folks! This list is great, but it's way too small. As @Spinster: says "Yes to more matching with datasets that describe art and culture terminology. Wikidata indeed still misses a lot of terms for styles and cultures, I have noticed that".
SQID shows this number of instances:
- 396 cultural movement (Q2198855): with subclasses
- 365 art movement (Q968159): with its 4 subclasses
- 300 art movement (Q968159): only direct instances
For comparison, AAT has 5569 Periods and Styles, using this query
select * { ?x gvp:broaderExtended aat:300264088; gvp:prefLabelGVP/xl:literalForm ?label; gvp:parentStringAbbrev ?parents }
Furthermore, there are concepts outside the Periods and Styles hierarchy (object types or techniques) that WD also terms Styles. Eg 300041507 nishiki-e. Mix-n-Match knows about this one:
- AAT:nishiki-e color woodcuts, woodcuts (prints), relief prints, <prints by process: transfer method>, <prints by process or technique>, prints (visual works), <visual works by material or technique>, visual works (works), Visual Works (Hierarchy Name), Visual and Verba
But why is it "Not matched"?
The recently approved Wikidata:Property_proposal/British_Museum_thesauri includes more thesauri that should be matched:
- Definitely these: 6070
- thes:ethname 3351
- thes:matcult 1483
- thes:school 482
- thes:ware 666 (ceramics ware)
- thes:escape 88 (clock escapement)
- Probably these:
- thes:technique 598
- Maybe even these:
- thes:nationality 250
- thes:political-state 191
- thes:state 263
Periods and Styles info is quite crucial for art research.
I've had a dream of unifying AAT and BM styles, and matching to Wikidata.
But it's a huge lot of work, not least because Wikidata is missing a lot of AAT Periods and Styles, but often has them as the corresponding Peoples, or Place, or an archaeological site, or language, or even individual person.
- Eg at AAT constraint violations:
- 300016563 Sotho (culture or style) -> Sesotho (Q34340) language, Sotho–Tswana peoples (Q7563990)
- 300020101 Hellenistic (Ancient Greek (culture or style)): Hellenistic art (Q221375), Hellenistic period (Q428995), Hellenistic civilization (Q913229)
- Eg at Wikidata:Property_proposal/British_Museum_thesauri
- Alhambra (Q47476) <- x56901 Alhambra style: style named after a place
- Achille Brocot (Q4673700) <- x86193 Brocot watch escapement: named after its inventor
AAT has over 45k concepts. Wikidatians have matched 8477 on Mix-n-Match, and I just posted 3324 potential matches (need your help!). But that's still less than a quarter, and even fewer Styles are probably matched.
So how can we move this forward? --Vladimir Alexiev (talk) 16:06, 7 March 2017 (UTC)
Discussion edit
- @Vladimir Alexiev: Plan B -- if you think you have got most of the potential matches that are properly legitimate and not constraint violations -- would be just to create the remaining items for either AAT or BM, with as much structure as you can gain from those databases, and then match from the other database to the items now in place on wikidata.
- Then do searches to try to maximise the number of named after (P138) connections you can make with similarly-named items. Eg a style might be named for a people, a language might be named for a people. Then, with a SPARQL query looking at all the things joined by recursive P138s, analyse the kinds of things found in those bundles -- perhaps into a table to show what other things are associated. This may help discussion as to whether there are particular types of things that should be merged, or perhaps have key sitelinks moved from one item to another. Jheald (talk) 16:28, 7 March 2017 (UTC)
- @Jheald: I don't think we have all potential matches (my 3.5k come from older Wordnet/Babelnet mapping, they're not complete). And adding all en-masse may not a good idea. --Vladimir Alexiev (talk) 18:27, 7 March 2017 (UTC)
- Something else that can be a good idea is to include subject named as (P1810) as qualifier on the external IDs, to indicate the preferred name for the thing in that database. (still called "credited as" in some languages, but moving towards "named as" to encourage wider use.) -- this makes it easier to spot potential bad matches when browsing items, or to systematically examine comparisons with queries, for particular subgroups. Jheald (talk) 16:35, 7 March 2017 (UTC)
- Agree, since AAT prefLabels are in plural. Getty CONA has similar needs, see this diagram.
- In the very beginning I tried to move forward the AAT matching. I don't have that much time right no, but I will try to help in whatever way! :-) --Marsupium (talk) 17:31, 7 March 2017 (UTC)
- Thanks, now I remember I've looked at this page. Can someone make complete navigation to the sub-pages, else people like me will miss them? @Marsupium: This is very thoughtful: are all of your matches reflected in WD? I'll ping the Getty about creating items for this section --Vladimir Alexiev (talk) 18:27, 7 March 2017 (UTC)