Wikidata talk:WikiProject Fashion

From Wikidata
Jump to navigation Jump to search

Home

 

Discuss

 

Items & Properties

 

Taxonomy

 

To Do

 

Specific items

 

Tools & Tips

 

Ask questions or share ideas about this project here!

Origins[edit]

This project grew out of a discussion of GLAM ontologies in the Facebook Wikidata + GLAM group. It seems there is sufficient interest to kick off a project. - PKM (talk) 21:19, 28 February 2017 (UTC)[reply]

Ontologies and vocabularies[edit]

This discussion has been moved to the Taxonomy tab.

Named fashion "looks"[edit]

What is Dior's New Look (Q2030942) an <instance of>? Thoughts? - PKM (talk) 01:44, 3 March 2017 (UTC)[reply]

I've experimented and put it as an instance of fashion collection (Q5480329), which I think may be right, unless you mean the "New Look" not as the 1947 Dior collection, but as a broader fashion trend.--Pharos (talk) 14:42, 6 March 2017 (UTC)[reply]
I think that works. - PKM (talk) 20:01, 6 March 2017 (UTC)[reply]

Fashion-related property proposal[edit]

WikiProject Fashion participants may be interested in this proposal, and ensuring such a potential property is framed correctly: Wikidata:Property proposal/wears.--Pharos (talk) 14:48, 6 March 2017 (UTC)[reply]

@Pharos: Thanks for that heads-up - PKM (talk)

Describing clothing in a museum[edit]

I have taken a stab at seeing how thoroughly we can describe an article of costume in museum collection. I'd love some feedback on Layton jacket (Q6759619) (take a look at this is in Reasonator, as well). Are there better properties to use?

I am particularly challenged by the coral silk taffeta lining. Two nested iterations of <material used> offends my sense of order. Would it be better to create an item for "silk taffeta" <subclass of> taffeta (Q909391) <material used> silk (Q37681)? This implies creating items for every fabric/fibre combination we want to use, but the more I think about it, it think it's the way to go - and it addresses Wikipedia's (and thence Wikidata's) tendency to cover the fibre, its yarns, and fabrics made from it all in one article. - PKM (talk) 18:35, 6 March 2017 (UTC)[reply]

personnalité du monde de la mode[edit]

Would "fashion icon" be a good idiomatic English translation of "personnalité du monde de la mode"? - PKM (talk) 04:37, 7 March 2017 (UTC)[reply]

My French is pretty basic, but after a little research and seeing the individuals it's applied to, this makes sense to me.--Pharos (talk) 17:10, 7 March 2017 (UTC)[reply]
Great, done. fashion person (Q27943370). - PKM (talk) 19:31, 7 March 2017 (UTC)[reply]

Showcase items[edit]

As a project goal, I'd like us to develop a handful of showcase items - Wikidata items that can serve as examples of best practices for the fashion project. These items should have rich metadata and be well-referenced. Possibilities are iconic designers, classic garments, influential brands, and important museum costume pieces. - PKM (talk) 20:30, 8 March 2017 (UTC)[reply]

Suggest showcase items to develop here!

Business of fashion, professions, occupations[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

We have two new sections on our Items and properties page.

  • Can you help build out best practice properties for the business of fashion?
  • Can you expand or suggest improvements to our list of professions, occupations, and crafts? (See Wikidata:WikiProject Occupations and professions for more info on professions and occupations ... it still seems pretty fuzzy to me.)

Comments and discussions on these items can go here on this page. Thanks! - PKM (talk) 03:32, 10 March 2017 (UTC)[reply]

History of clothing, costume, dress, and fashion[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

Sorting out history of fashion, history of clothing, history of costume, history of dress is fraught. WP articles on substantially similar things in different periods use variously "costume" or "mode" in French, and "clothing", "fashion", or "dress" in English. I've been struggling with how to use these terms on EN WP for over ten years. It's not all going to all fall neatly into place now. :-)

At the moment, I am working with these parameters:

I suggest using the structure in Tortora's textbook Survey of Historic Costume:A History of Western Dress which will give us a reference to cite for the structure. She has Ancient World > Middle Ages > Renaissance > Baroque & Rococo > Nineteenth Century > Twentieth Century. This will require very few changes to what I've started.

We'll need to find a reference and an expert on Asian dress to build out that structure. (Do you know someone?)

Please, I need your feedback on this approach. Do you have another idea? Do you hate parts of this? - PKM (talk) 21:18, 14 March 2017 (UTC)[reply]

Further thoughts[edit]

Looking at some similar items, I think my schema is mixing concepts/field of study and "things". So for example, currently clothing in ancient Greece (Q522648) has these subclasses which are individual garments. I don't think garments can be a subclass of history.

Most of these "clothing of [period]" items have subclasses that are types of garments and links to articles in many wikipedias. For sanity's sake, I think these should remain subclasses of clothing but not be subclasses of history of costume. (I note that "renaissance art", "ancient Greek art" etc. are not subclasses or part of "art history".)

And here are some some examples of similar item structures in other domains.. - PKM (talk) 02:18, 15 March 2017 (UTC)[reply]

Europeana Fashion ID[edit]

FYI: As originally proposed by user:Thierry Caro, the Europeana Fashion creator ID (P3482) property has now been created (documentation at: Property talk:P3482]).
user:Spinster has added it to Mix'n'match and we've got it up to over 40% matched very quickly: https://tools.wmflabs.org/mix-n-match/#/catalog/397 :-) Wittylama (talk) 13:02, 23 March 2017 (UTC)[reply]

Followup: many of the remaining items are people's names, but is common for the fashion label/fashion house (i.e. the brand) to be named directly after the person who is the designer/founder of the company. E.g. Agnès Troublé (Q2827013) is a person but Agnès b. (Q282967) is a company. I've been tagging these to the company/brand wikidata item (when possible) as I feel this is what is represented by the Europeana ID more specifically in those cases. Agreed Spinster, PKM et. al?? Wittylama (talk) 13:30, 23 March 2017 (UTC)[reply]
Yes, absolutely. Spinster 💬 15:01, 23 March 2017 (UTC)[reply]
Agreed. - PKM (talk) 18:56, 23 March 2017 (UTC)[reply]

Another Europeana Fashion ID: Vocabulary[edit]

I have posted a Proposal to add a property for Europeana Fashion Vocabulary ID. Can someone add the links for the examples and the formatter ID for me?

Also created Europeana Fashion Vocabulary (Q29016777): multilingual controlled vocabulary for fashion and costume.

@Wittylama, Thierry_Caro, Spinster, Pharos: FYI for your comments/suggestions. - PKM (talk)

Master work lists[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

I've created a set of master work lists of items related to this project. Find them on the To Do tab. - PKM (talk) 20:42, 24 April 2017 (UTC)[reply]

Folk dress, national costume, etc.[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

This is a "best practices" question. Currently, there is no standard for the English labels for articles on folk costume. We have options:

  • folk dress of X
  • folk costume of X
  • traditional dress of X
  • traditional costume of X
  • traditional clothing of X
  • national dress of X
  • national costume of X
  • clothing of X
  • X-ian folk dress
  • X-ian national dress
  • X-ian costume
  • X-ian clothing
  • traditional X-ian clothing
  • traditional X-ian costume

... and more (here's the current list). All of these constructions are standard in English. Someone might search for any of them. My question is, from the perpective of the usefulness of our database, should we have a single standard for these labels in English (with the caveat that not all traditional clothing is thought of as "national dress")? Should all of these options appear as aliases? Should I stop worrying about it? - PKM (talk) 03:02, 3 June 2017 (UTC)[reply]

You need to think in the wikidata way. So not "Mayor of London" but "occupation=politician" and "position held=Mayor" (with administrative body as qualifier and start and end dates if you have them). So in this case, one item for "traditional costume" and then we need to define qualifiers for region, time period and part (headdress/skirt/robe/boots etc) --just my 2c. Jane023 (talk) 07:08, 3 June 2017 (UTC)[reply]
We can certainly do that in statements - we have traditional costume (Q3172759): expression of identity through costume, which is usually associated with a geographic area or a period of time in history - but it doesn't help with the dozens (hundreds?) of items that are generated by individual Wikipedia articles on "folk dress of X". These all need labels. Further research suggests anyone doing a manual search will find things if the keywords are in the label or the description - so if the label is "Welsh folk costume" and the description is "traditional rural clothing of women in Wales, adopted as national dress" then I think a search for any of the variations I listed above would find the item. I think I'll try that with one of the un-translated items and see what happens. And if there's an EN wiki article, then whatever that is called should probably be an alias as well. I'd just like to avoid 12 or 13 long-winded aliases for every item.
And yes, my theory of search is correct. As long as "folk", "national", "traditional", "dress", and "costume" are all in the EN label and description somewhere, along with the demonym and country/region name, then any of the seacrh strings represented by my long list of variant titles above will return the correct item. Now I just need to standardize my translations. - PKM (talk) 20:54, 3 June 2017 (UTC)[reply]
@Jane023:, @Spinster: From your perspectives, do you have a preference for "folk dress", "folk costume", or "traditional dress" for the default English label? - PKM (talk) 20:28, 3 June 2017 (UTC)[reply]
I prefer traditional dress, since there is more traditional dress styles than "folk" (e.g. a blue sash for some royal families could be a form of traditional dress for special occasions). If you set up the data modelling correctly then the label won't matter so much because you can search with the query service. Jane023 (talk) 21:14, 3 June 2017 (UTC)[reply]

Iconclass Notation[edit]

Iconclass notation (P1256) includes IDs for some clothing items. Here are some non-literal matches I have made that may require further research and discussion at some point:

- PKM (talk) 21:48, 18 November 2017 (UTC)[reply]

Ontology and state of the project[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

As the Fashion project approaches its first anniversary, I think it's a good time to assess our progress. These are my personal thoughts.

In general[edit]

I am generally quite happy with the state of our classifications of clothing (Q11460) items. To date, 1024 items have been matched to a Europeana Fashion Vocabulary ID (P3832). The lower levels of our classification system - where we say things like "a poodle skirt is a <subclass of> skirt" - are pretty solid, and on their way to well-referenced. There is still much work to be done on military uniforms and armor, especially Japanese armor. And there are some kludges and hacks in the classification that were made to get everything linked and findable. These still need work.

Ontology[edit]

I am unhappy with the upper ontology. As I recently pointed out on Project chat, here are comparative hierarchies for "clothing" in Wikidata and AAT:

Wikidata: entity > object (philosophy) > abstract object > concept > result > goods > product > clothing

Getty AAT: object > furnishings & equipment (hierarchy name) > costume (hierarchy name) > costume (mode of fashion) > clothing

Clearly, not all clothes are products, and goods are not abstract objects. There may be other issues that I am not aware of, since I spend most of my time working very far down in the weeds on a daily basis.

I am pinging the Ontology project here, and asking for a look specifically at the classification of clothing (Q11460) to see if they can suggest improvements. (@TomT0m:, thanks for the suggestion!)

Specific questions about classes[edit]

Some other specific questions I have:

    • Should there be a direct relationship between "toga" and "clothing in Ancient Rome"?
    • Is "clothing in Ancient Rome" a class of garments, or a field of study, or an aspect of Ancient Roman culture, or a mix of these?
  • Are the early distinctions made between "fashion", "costume", "clothing" and "dress" reasonable from an ontological perspective and consistent with the way other areas are mapped in Wikidata?

Please let me know what you think, and add any questions you may have! - PKM (talk) 21:13, 8 February 2018 (UTC)[reply]

AAT[edit]

@PKM: Just as an aside first, before the ontological discussion, I've put up an outline of the AAT tree (first four levels only), and also the full tree here. For the costume parts, use 'find' in your browser to go to "Costume (hierarchy name)" (which is under 'furnishings and equipment'), and "costume components", which is under "Objects -> Components". Entries with matches to Wikidata are blue-linked. (Data as of about a week ago). It looks like the tree has been worked quite hard already, but I was planning to have a go in the next few days with OpenRefine, to see if it could match some more. Jheald (talk) 21:59, 17 February 2018 (UTC)[reply]

@Jheald: Very cool. It would be great to have a fresh pair of eyes on this. I'm sure there are clothing items imported from the various Wikipedias with no statements as yet. I got a bit burned out on going through Wikipedia categories in various languages looking for bits of folk costume and whatnot. The worklists on this page have AAT IDs if that's any help to you. - PKM (talk) 22:23, 17 February 2018 (UTC)[reply]
@PKM: I had a go with OpenRefine. Quite clunky, and not as efficient as MnM; but I was able to give it just a portion of the AAT tree to work with (ie the "costume (hierarchy)" and "costume components" parts), and I did get about 30 or so new matches. (contributions). I hope to do a bit more analysis with some of the AAT-matched items in the next few days. JhealdBatch (talk) 17:57, 22 February 2018 (UTC)[reply]
Great! Good catches, all. I’m deep in the weeds with my textiles/fibers sorting project this week, but I’ll need a break from that soon (polymers, ugh!). - PKM (talk) 20:15, 22 February 2018 (UTC)[reply]
As an experiment, I have now added part of (P361) as a qualifier to entries in the fashion part of the Art & Architecture Thesaurus ID (P1014), pointing to the next matched item upward in the hierarchy. Use of part of (P361) isn't ideal -- I still think that a new specific property for this kind of relation would be better (see Wikidata:Property proposal/broader concept) -- but it gives a chance to experiment.
So for example, this query tinyurl.com/yctm5t2k tinyurl.com/y7oec9sp pulls out all items matched to the AAT tree below costume accessory (Q1065579), and lists them if they do not have a subclass of (P279) relationship corresponding to the relationship in the AAT tree.
It can pull out a number of issues, eg the wrong kind of item matched to an AAT id; items in the tree with no subclass of (P279) at all; or items in the tree missing what might be an appropriate extra subclass of (P279) connection.
Unfortunately the query runs much much more slowly than I had been expecting -- taking over 20 seconds of processor time, just to look at this part of the tree. Setting the same query to run on all of the items under costume (Q9053464) it times out, taking over 50 seconds just to find the items, even though they are arranged in a simple tree of direct hierarchical links. I have asked at Request a Query whether anyone has any ideas as to why this is taking quite so long, and whether they have any ideas to speed it up. (I have now found a workaround for this, and the query is now acceptably quick; though I still don't understand why the original was so slow).
Nevertheless, I hope this may still be quite interesting.
For comparison and reference, I have put up the tree of the fashion part of the AAT hierarchy at Wikidata:WikiProject_Fashion/Taxonomy/aat.
In adding the links I have also added qualifiers subject named as (P1810) to give AAT's preferred name for the matched class, and sourcing circumstances (P1480) = hierarchical link is not direct (Q50095342) if the nearest upward item we can match is not the direct one, but one from further up the hierarchy.
This is really only intended as an experiment, and I'm quite happy to back any or all of it out again, but I'd be interested to know what people think. Jheald (talk) 14:18, 25 February 2018 (UTC)[reply]
@PKM: Here's now also a similar query for the Europeana Fashion Vocabulary tinyurl.com/yaag47tb. Again, the first column gives a item that's been matched to the EFV, the higher column gives the nearest match upward in the EFV hierarchy, while the class column in the middle gives the classes that the item is part of, none of which are subclasses of the upward match.
Also, a tree for the EFV hierarchy, Wikidata:WikiProject_Fashion/Taxonomy/efv, similar to the one I put up for the AAT hierarchy -- though maybe you have something like this already, it seems impressively thoroughly filled in!
Looking at some of the rows in the query, some the item simply doesn't have any subclass of (P279) relationship stated currently. In quite a number of other cases, I found the "higher" class is actually a subclass of the class that the item is currently stated as subclass of (P279). (tinyurl.com/ya9x96lw; remainder tinyurl.com/ya35olk4). Some of the time, it appeared that could be a sign that the subclass of (P279) could be refined to a lower-level, more direct intermediate. But in other cases I'm not so sure.
In these other cases, it seemed to me that while in the context of fashion it might be true that any/all items in <class A> would also be in <class higher>, that in a more general context that might not necessarily be the case -- and our broader classing probably makes sense. It strikes me it might be worth thinking about an additional qualifier / values to record a reason why <class A> should not be subclass of (P279) <class higher> -- eg if the statement would only necessarily apply in the application domain of the thesaurus, or if it's a different property other than P279 that better expresses the relationship -- so that such cases could be excluded from the query. Jheald (talk) 22:18, 28 February 2018 (UTC)[reply]
@Jheald: "it seems impressively thoroughly filled in" - thank you! I have the EFV vocabulary as a spreadsheet and i've been working through it for about a year. I am also in touch with the EFV team about some open questions.
There are many cases in their vocabulary where the "broader" assignments make no sense outside of their own vocabulary - especially for items other than the clothing hierarchy (an example that comes to mind is "cocktails" broader = "fashion event").
Some sort of record of why an editor chose a different parent class would be helpful. We can be so much more granular than either EFV or AAT.
Thanks so much for doing this. I want to spend some time going through your points and documents, but it may be another few days before I can focus on that. - PKM (talk) 22:59, 28 February 2018 (UTC)[reply]
@Jheald: Do your trees automagically update as things are added/matched? - PKM (talk) 02:23, 1 March 2018 (UTC)[reply]
You mean, like a Listeria query? No, not as yet. Rebuilding the tree pages means running a WDQS query to get an updated list of what's been matched, then re-running a script to rebuild the page. But that certainly could be automated, if it was useful.
The WDQS queries above are of course up to date to within about a minute, as to subclass of (P279) statements that items have on them, so they will change if/as those P279s are edited.
The part of (P361) qualifiers to reflect the external hierarchy won't update themselves if it was an indirect relationship and a new intervening member gets matched, so at any time there may be some recent detail that they are missing. That's probably not so big a problem, because the broad hierarchy will still be right. Again, with a script it should be easy enough to identify ones to update; in principle one could get bot permission to get that script to re-run as regularly as desired. I am not 100% sure how a bot passes statements on to QuickStatements (or maybe it just re-writes the statements itself), this isn't something I have yet done. But I think it should just be a question of working out how to do the appropriate plumbing. Jheald (talk) 11:18, 1 March 2018 (UTC)[reply]
@Jheald: Thanks for the info. I've matched a bunch of missing items in both hierarchies, and I am working through the AAT query list now. The problem with keffiyeh (Q463298) is actually a mucky problem that's a result of "kerchief" and "bandana" being munched together in EN Wiki. I'm going to have to tackle that soon. I suspect it has ghosts in multiple places. :-)
I disagree with AAT making what is essentially a class for the Orkney hood (Q21051423), which is a single archaeological find with no other exemplars. - PKM (talk) 22:11, 3 March 2018 (UTC)[reply]
@PKM: Well, if the query is exposing tricky issues, it's probably doing its job :-) Let me know if/when you'd like anything updated; also, any thoughts on the process of comparison with external hierarchies and what could be done to improve it, now you've had a chance to go through in detail a bit. Jheald (talk) 10:09, 4 March 2018 (UTC)[reply]

──────────────────────────────────────────────────────────────────────────────────────────────────── @Jheald: The query is definitely doing its job! It's great to see a group of related inconsistencies side-by-side and then decide how to harmonize them. I wasn't sure how practical the "broader concept" qualifier would be, but it's actually been incredibly useful - so thank you for that! I'll let you know if I think of any additional tools like this that would be helpful. - PKM (talk) 20:03, 4 March 2018 (UTC)[reply]

@PKM: Just to let you know, I have now updated AAT and EFV items to use broader concept (P4900) for the hierarchical relation, rather than part of (P361). The revised queries are now tinyurl.com/yaephu2l for the EFV tree, tinyurl.com/y79z79zj for where the EFV higher item is a subclass of our current subclass, tinyurl.com/yac85yg3 for where it isn't (but occasionally should be); tinyurl.com/yaev7okd for the AAT tree below costume accessory (Q1065579); and tinyurl.com/ya4cotlj for the broader AAT hierarchy below costume (Q9053464). I've also updated the pages with the extracts of the AAT and EFV trees. Jheald (talk) 14:04, 5 March 2018 (UTC)[reply]
@Jheald: smock (Q12338240) has two parents in AAT. I have added a second "broader concept". It doesn't seem to have broken your query, so I'll do some more. :-) If we put a single value constraint on broader concept (P4900), these will all need to be exceptions, but I can handle that. - PKM (talk) 01:18, 6 March 2018 (UTC)[reply]
@PKM: My script should have caught that. I don't know why it missed it. Possibly when I extracted the AAT hierarchy, I only extracted the principal parent, not additional parents. Let me look into that tomorrow, and see whether I can add all of them systematically. Jheald (talk) 01:29, 6 March 2018 (UTC)[reply]
I have only ever thought of a painter's smock -- didn't know the word had an earlier meaning as a chemise-like undergarment. Intrigued that the Danish sitelink is "Særk" -- presumably cognate to "sark" in Scots, as in Robert Burns's famous "cutty sark" (roughly = "shortie nightdress"?) Jheald (talk) 01:35, 6 March 2018 (UTC)[reply]
@Jheald: I'm tempted to split these two senses of smock, AAT or no. I think the idea that the earlier smock "protects" your (not washable) outer garments from sweat is a dubious link to say the undershirt sense and the painter's smock sense are the same concept. But I'll let it ride - I'm not sure how many languages use the same word for both senses anyway. Throw in chemise (Q1604733) which was called a "smock" in English until the 18th century - not to mention smock (Q1743902) - and the whole thing gets truly muddled! Ah well, we are not called on to structure the universe, but to record how others have structured it. Or so I keep reminding myself. It's not like costume historians all agree on any of this either. - PKM (talk) 01:56, 6 March 2018 (UTC)[reply]
And yes, you're right about sark. A northern word for "shirt" (and also for chainmail shirt). - PKM (talk) 01:58, 6 March 2018 (UTC)[reply]
"Broader concepts" from the "Additional parents" classification on the Getty pages are now added; and I've also updated the hierarchy page at Wikidata:WikiProject_Fashion/Taxonomy/aat to include them. The only slight wrinkle to note is that the hierarchical link is not direct (Q50095342) marker now might apply to one or other or both of the parent links -- short of creating a whole new AAT statement for the same ID value (which QuickStatements doesn't really do), I can't think of a way to attach it to just one or the other of the parent links. Thanks for spotting this. Jheald (talk) 17:47, 6 March 2018 (UTC)[reply]
Another query: tinyurl.com/ycxpe82q These are items below costume (Q9053464) where we have more than one AAT entry corresponding to a single item here. The last column indicates if there is a child/parent relationship between the two at AAT. Jheald (talk) 18:05, 6 March 2018 (UTC)[reply]
Another query that may be useful: tinyurl.com/yahoyenutinyurl.com/y6wcmrjy Here the first item is in the subclass tree of the second item, and both have AAT IDs; but the first item is not below the second item in the AAT hierarchy. Instead, the third item gives the next highest match up the AAT hierarchy (which is also not below the second item, even though AAT knows about both of them).
Could use some further development, to only show the highest subclass out of a tree of subclasses for which this is true. I think this is now done, though the query could possibly find some further speed tweaks. In some cases it looks as if our subclass/superclass relationship is a bit odd; in other cases it looks as if it is AAT that is missing a connection. Jheald (talk) 14:20, 8 March 2018 (UTC)[reply]
Update: faster version: tinyurl.com/ycoaj9dx Jheald (talk) 13:10, 9 March 2018 (UTC)[reply]
Equivalent query for EFV: tinyurl.com/yc5ex9lq Only real divergence seems to be that our notion of costume (Q1410477) (theatrical sense) seems to be a bit wider than theirs. Jheald (talk) 13:00, 9 March 2018 (UTC)[reply]
@Jheald: Still slogging through these, disentangling, reclassifying, etc. It's taking time since I'm generally improving items as I go. There are definitely some AAT parent-child relationships that don't meet the qualification that "all instances of [child] are instances of [parent]" (example: featherwork (Q2864716) and costume accessory (Q1065579)).
I do see a future "problem": if an editor creates a new Qitem for a "missing" level in the AAT hierarchy, they won't know to fix the broader concept (P4900) for all the existing child items which previously skipped this level. Can a bot keep these current? - PKM (talk) 23:32, 11 March 2018 (UTC)[reply]
@PKM: Hope you haven't minded too much all these being dropped in your lap. (And is a dinner jacket really a 'coat'? I do wonder about some of the AAT's classifications sometimes!)
You're quite right, it would be good to have a script that would identify updates to the broader concept (P4900). I'll try and get something written. A complicating factor is that Quick Statements isn't very good at just changing a qualifier -- updating it with QS would require the entire statement to be rewritten including all qualifiers, all dates, all references etc, which a) makes a bit of a mess in the history, and b) means one has to be quite careful not to miss anything in the extraction. So such updates might be better done by a bot; so I may need to think how to do that. On the up-side, the hierarchy as a whole still works, even though it is now missing a new node that ought to be included; so even without that node, most queries will still make sense, unless they're really focussing on comparing the local hierarchies. I'm also a bit tied up at the moment with a project on Commons that has been hanging around too long, and now needs some urgent attention. But I'll see what I can do. Jheald (talk) 23:44, 11 March 2018 (UTC)[reply]
@Jheald: Thanks for thinking about ways to address this, and no rush at all. It's going to be a while before this stuff is all cleaned up (if ever!). I'm going to need to take a break from it myself. - PKM (talk) 23:52, 11 March 2018 (UTC)[reply]
@PKM: Obviously, look after yourself, that needs to be first and foremost. But what you have been quietly building up here is properly impressive, also all the work on Commons that's been part of the project. Jheald (talk) 23:59, 11 March 2018 (UTC)[reply]
Jheald Getty AAT has a circular relationship between 300212545 'body armor for the head' and 300036794 'helmet (protective wear)'. Both concepts are covered in most Wikipedia articles. 300212545 is a placeholder guide level in AAT, but there is no other parent category for various ancient and medieval helmets. I'm going to leave the helmrt hierarchy alone except for reserving "combat helmet" for 20th century and forward military uses. - PKM (talk) 19:18, 15 March 2018 (UT

Clothing and Costume[edit]

@Infovarius, ArthurPSmith, jheald:

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

Reacting to recent edits, reversions, and edit comments oon clothing (Q11460) and costume (Q9053464) (and bringin in dress (Q28955655) as well): I have tried to distinguish between these concepts as:

Scholarly sources in English use "costume" and "dress" in all three senses, which doesn't help. The assignment "clothing" <subclass of> "costume" (which Inforvarious has removed) was a convenient hack while I was trying to sort everything out, and I have been dubious about it for some time.

If we think of "clothing" as individual classes of objects, then <subclass of> "costume" may be incorrect. As I have said before (upstream), the upper ontology for this project is strange and tricky, but no one seems to be particularly interested in working on that right now (which is okay; it's not like biomedicine, fixing it won't save lives).

Anyway, how does everyone feel about "clothing" as <subclass of> "costume"? I am tempted to use a <part of> relationship here instead, but I'd like others to weigh in. Getty AAT, being a vocabulary not an ontology, can logically support either relationship. - PKM (talk) 20:14, 8 August 2018 (UTC)[reply]

<part of> is fine with me, I agree the AAT supports some sort of broader relationship but it's not specific about what! ArthurPSmith (talk) 20:21, 8 August 2018 (UTC)[reply]
Hmm - also maybe costume (Q9053464) should be split up - some of the linked wiki articles talk about in a general sense, but some focus on the specific "dress-up" meaning, of putting on a style to pretend to be somebody else. ArthurPSmith (talk) 20:27, 8 August 2018 (UTC)[reply]
There are items for these - fancy dress costume (Q29403086) and costume (Q1410477). Just need to map the articles. - PKM (talk) 20:34, 8 August 2018 (UTC)[reply]

No. Costume, at least in English, comes with a certain meaning of completeness. If you are describing a golden sword found in some archeological dig somewhere in the world, then it is of course "instance of sword" but you might be more specific, such as a ceremonial sword that was part of some specific costume. When it comes to clothing, a single brooch would be "instance of jewelry", and could be construed as a clothing element, but I would be reluctant to call it part of a costume. So I think clothing is broader than costume. I agree with your ontological concerns - hasn't this been done before? What does the AAT say? Meanwhile, check out this Wikidata:WikiProject Women/Portraits of women. Should I recreate it for men as well? We also have this, which includes lots of wedding clothes Wikidata:WikiProject sum of all paintings/Pendant portraits. Jane023 (talk) 06:03, 9 August 2018 (UTC)[reply]

@Jane023: The AAT hierarchy can be found here (note that it's a while since the page was created, so the matches to Wikidata may be slightly out of date -- User:PKM has been busy!) As you can see, AAT makes "costume" its highest-level term for this sector. For what it's worth the Online OED gives as definition 2a: "The style of clothing, hairdressing, and personal adornment typical of a particular place, period, group, etc.; an example of this. Also: such styles of clothing, etc., as a subject of study." Presumably the AAT is making "costume" its parent term for all 'items of costume' or 'aspects of costume'. The exact sense in which AAT intends its term can be read at [1]. Jheald (talk) 20:54, 9 August 2018 (UTC)[reply]
There may be yet another concept called "costume" in play here. "An afternoon costume by Patou" would indicate a dress or suit, perhaps with associated accessories, but that's a fourth meaning of "costume" from costume (Q9053464), fancy dress costume (Q29403086) and costume (Q1410477) - perhaps an alias of both ensemble (Q29353484) and dress (Q200539) (frock, gown). - PKM (talk) 21:02, 9 August 2018 (UTC)[reply]

"Evening" garments[edit]

I've recently run into a categorization at The Met Museum which has distinct object names for "evening" wear. Specifically, these are popular:

  • evening dress
  • evening blouse
  • evening suit
  • evening ensemble

We don't currently have separate Wikidata items for all of these, though we seem to follow AAT in that "evening dress" is prominent enough to have evening dress (Q2144456) while the others do not. The question is, how should we model these? Should we simply make items instances of the main item like blouse (Q152563) and then also evening wear (Q1908472)? I'm also including a long-ish list of other evening items from The Met and their frequency. -- Fuzheado (talk) 15:01, 23 August 2019 (UTC)[reply]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

Evening dress                                          1864
Evening ensemble                                        641
Evening shoes                                           271
Evening coat                                            245
Evening pumps                                           100
Evening cape                                             85
Evening jacket                                           72
Evening suit                                             63
Evening bag                                              63
Evening slippers                                         61
Evening gloves                                           48
Evening sandals                                          45
Evening stole                                            40
Evening hat                                              37
Evening wrap                                             35
Evening blouse                                           35
Evening purse                                            34
Evening skirt                                            32
Evening belt                                             25
Evening boots                                            22
Evening shawl                                            22
Evening clutch                                           20
Evening vest                                             18
Evening pouch                                            18
Evening jumpsuit                                         16
Evening bonnet                                           13
Evening shirt                                            12
Evening mitts                                            11
Evening pantsuit                                         10
Evening overblouse       9
Evening trousers         9
Evening bow tie          9
Evening top hat          7
Evening headdress        7
Evening muff             7
Evening bodice           6
Evening sweater          6
Evening pajamas          6
Evening suspenders       5
Evening scarf            5
Evening bolero                                         5
Evening cravat                                         5
Evening collar                                         5
Evening apron                                          4
Evening cloche                                         4
Evening robe                                           4
Dress, evening                                         4
Evening cap                                            4
Evening cloak                                          4
Evening shoulder bag                                   3
Evening overdress                                      3
Evening dickey                                         3
Evening poncho                                         3
Evening toque                                          3
Evening tunic                                          3
Evening oxfords                                        3
Evening mules                                          3
Evening caftan                                         3
Evening stockings                                      3
Evening turban                                         2
Evening snood                                          2
Evening fan                                            2
Evening slip                                           2
Evening ascot                                          2
Evening dolman                                         2
Evening sweater set                                    2
Evening sash                                           2
Evening accessory set                                  2
Evening mantle                                         2
Evening underdress                                     1
Evening jabot                                          1
Belt, Evening (Cinch Belt) (Women's)                   1
Evening bodysuit                                       1
Evening cheongsam                                      1
Evening veil                                           1
Evening pillbox                                        1
Vest, Evening                                          1
Evening overskirt                                      1
Evening short shorts                                   1
Evening moccasins                                      1
Evening teddy                                          1
Evening drawstring bag                                 1
Evening belt buckle                                    1
Evening minaudière                                     1
Evening waistcoat                                      1
Evening tam-o’-shanter                                 1
Evening thongs sandals                                 1
Evening coatdress                                      1
Evening trench coat                                    1
Evening headscarf                                      1
Evening fichu                                          1
After a very rapid look, I could say that for me Evening wear should be the main category, as is related to the formal evening occasions (parties, ceremonies etc). And here we can have some others sub-categories abito da sera (evening dress), elegant tailleur, smoking, frac etc. And then ceremony shoes, hats and accessories. I want to share with you other two site related to the fashion: Fashion heritage and Europeana Collections, both managed by Europeana. --Camelia (talk) 15:27, 23 August 2019 (UTC)[reply]
I do believe that the Nordic Museum have been looking a little bit on fashion. @Ambrosiani: Any ideas? Ainali (talk) 17:56, 23 August 2019 (UTC)[reply]
@Fuzheado, Camelia.boban: I'd have no issue with just using <instance of> the main item like blouse (Q152563) and also evening wear (Q1908472) for most of these, though we could certainly justify separate items for "evening blouse", "evening suit" and "evening ensemble" if you want to make them. We could also add "evening coat" as an alias to opera coat (Q7096642). The enwiki article could use expansion (it's called "opera cloak" but then discusses opera "coats"), and I've made a minimal WD item for it. Perhaps flag it for the editathon? - PKM (talk) 21:22, 23 August 2019 (UTC)[reply]

Hmm yes, this is probably a "genre" of clothing in the "clothing as art" sense. Of course even less glamorous clothing is still art in the sense it can be collected (because someone wore it or whatever), so we need "casual Friday", "mourning dress" etc. Jane023 (talk) 06:21, 24 August 2019 (UTC)[reply]

Thanks for the great insights! For now I might do it both ways (ack) to add it as instance of both the main type and "evening wear" while also adding genre as "evening wear" using the principle of prioritizing usefulness and discoverability over technical strictness. Let me know if folks have any major objections. -- Fuzheado (talk) 16:59, 3 September 2019 (UTC)[reply]

Seasons for fashion - fall/winter 1993–94[edit]

Has anyone modeled how to record the season for garments? At The Met, we have dates such as "fall/winter 1993–94" for dresses and ensembles, and I don't recall seeing this being done in Wikidata. Thanks! -- Fuzheado (talk) 17:03, 3 September 2019 (UTC)[reply]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

I don't think anyone has done this for specific garments.
We have an item season (Q29346471) (= fashion season) with subclasses autumn-winter (Q29346472) and spring-summer (Q29346473), all following the Europeana Fashion Vocabulary. Note that I made season (Q29346471) a subclass of both fashion collection (Q5480329) (following EFV) and season (Q25938183). I am not sure this is right for Wikidata; it leads to a situation where one could create a generic item "fall/winter 1993–94" <instance of> autumn-winter (Q29346472), and then "Ralph Lauren fall/winter 1993-94" could be either instance of (P31) or time period (P2348) = "fall/winter 1993–94". I am inclined to remove the <subclass of> "fashion collection" from "(fashion) season", so it just refers to time and we always have to use two separate statements:
Either way, once you create "fall/winter 1993–94" <instance of> autumn-winter (Q29346472), you should be able to attach it to a dress or ensemble using the "time period" property even without making a separate item for a specific designer's fall/winter 1993-94 collection. - PKM (talk) 21:21, 3 September 2019 (UTC)[reply]
I suppose one could also carry season and year as two separate fields, "time period" = "fall/winter" and "time period" = 1993-1994 one-year-period (Q53385494) or some such. - PKM (talk) 21:31, 3 September 2019 (UTC)[reply]
Somewhat related - I wonder how "universal" this season designation is. It seems very European and American, and is specific to northern hemisphere. If we are trying to figure out how to make Wikidata more universally useful, I do wonder whether the fashion industry in the southern hemisphere just follow the lead of the northern, or whether they have their own way of doing things. -- Fuzheado (talk) 15:09, 5 September 2019 (UTC)[reply]
@Fuzheado: Cape Town Fashion Week uses the same seasons, as does Vietnam International Fashion Week. In India Lakme Fashion Week has a "winter/festive" edition. I think these are pretty standard. - PKM (talk) 20:10, 5 September 2019 (UTC)[reply]

Modeling ready to wear items with four elements[edit]

Hi all, I'm pondering how to model ready-to-wear items that might have up to four legitimate elements:

  • desginer
  • design house
  • manufacturer
  • retailer/department store

As an example, see this Bill Blass suit from The Met [2] where we have three of the above. Let's say hypothetically for this exercise, there was also a "design house" specified. I'm thinking this is perhaps the right way to go:

The problem is we don't have a property that fits what a retailer or department store does. Maybe distributed by (P750)? But that's mainly for media and records, though the property proposal seems to indicate it can be used for more than that. But a department store is not just a "distributor" so that's not a perfect fit. Ideas welcome! -- Fuzheado (talk) 17:46, 5 September 2019 (UTC)[reply]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

I like your proposed model. For "retailer", if we want to use an existing property, we might use commissioned by (P88). Otherwise I think the best solution is to propose a new property "dealer" alias "retailer, seller, retail outlet, exclusively available through" meaning "organization or individual that offered the subject for sale". I'd be happy to propose the property if there's consensus that it's needed. - PKM (talk) 19:56, 5 September 2019 (UTC)[reply]

Property proposal "worn by"[edit]

I have proposed the new property "worn by" per discussion above. Please comment on the proposal if you have an opinion! - PKM (talk) 20:30, 11 September 2019 (UTC)[reply]

Fashion seasons (again) and modeling collaborations[edit]

I have created one fashion season Fall/Winter 2013-2014 (Q67327783) because I wanted to use it to model a collaboration between James Lock & Co. Ltd (Q67218668) and Vivienne Westwood as a significant event.

Is "collaboration" the right word here, or would we prefer "partnership"? Please look at what I did and let me know if we like it or want something different: Fall/Winter 2013-2014 (Q67327783). - PKM (talk) 20:57, 11 September 2019 (UTC)[reply]

CFDA Fashion Awards[edit]

PKM (talk) Wittylama (talk) Thierry Caro ( talk) Ainali (talk) Spinster 💬 13:02, 6 March 2017 (UTC) Pharos (talk) Jane023 (talk) 07:22, 18 March 2017 (UTC) Tris T7 TT me Fuzheado (talk) 11:28, 25 June 2019 (UTC) Camelia (talk) 04:51, 31 July 2019 (UTC) Arkirkland (talk) 14:48, 7 September 2019 (UTC) NettieLibrarian (talk) 15:46, 19 May 2020 (UTC) Shani Evenstein (talk) 21:58, 9 February 2022 (UTC) Samantha Levin, MSLIS (talk) 17:44, 8 February 2021 (UTC) Lukutroel (talk) 12:14, 06 November 2021 (UTC) Wallacegromit1 (talk) 13:20, 5 June 2022 (UTC)[reply]

Notified participants of WikiProject Fashion

I'm working on adding the Council of Fashion Designers of America Fashion Awards to Wikidata. Following the model used for the Academy Awards, there will be an item for each named award when I am done. The working list is here: Tabernacle CFDA Fashion Awards. You can add "award received" for these to fashion designer bios. (Let me know if the award you need hasn't been created yet and I'll get it done. There are a lot of named awards over 40 years.) - PKM (talk) 22:09, 19 September 2019 (UTC)[reply]

New Work Lists?[edit]

I'm thinking about adding some additional list pages for costume components, for example sleeves, or necklines. Would it be appropriate to add those under the Master Work Lists section on this project's To Do page? Perhaps at the level of the Clothing bullet, create a bullet point for Costume Components, and then nested under that each individual list as it is created?

Is adding a new list as simple as copying the template from the existing lists but changing the SPARQL query to reflect the class to be represented?

I have an ongoing project related to this that I haven't had much time for lately, but there's some talk on the forum for the Costume Society of America about how to describe objects for cataloging if you don't know the vocabulary, and where to find pictures to help with that, and I was hoping I could throw together some pages quickly to provide easy access to all the great resources here!

Please let me know what you think. Arkirkland (talk) 18:10, 3 November 2019 (UTC)[reply]

@Arkirkland: Your proposal sounds good. And yes, I always create these by copying the template from an existing list and changing the SPARQL query. You have to click the "manually update list" link after you create the new subpage (sometimes twice) to see the new items. Let me know if you get stuck. - PKM (talk) 20:22, 3 November 2019 (UTC)[reply]
@PKM: Great-thanks! I've added lists for necklines, collars, and sleeves so far. It definitely helps to see them like this, and where some things perhaps don't belong, or have images that aren't quite right. Hopefully I'll have time to continue with this soon. Arkirkland (talk) 02:24, 4 November 2019 (UTC)[reply]
@Arkirkland: these are really great! We don’t have “dolman sleeve” or “Bishop sleeve”. How is that possible? :-)
I’ve been trying to find a proper image of a 1940s sweetheart neckline that is licensed for Commons - no luck so far. Thanks for doing these. - PKM (talk) 06:01, 4 November 2019 (UTC)[reply]

More new work lists[edit]

Here's a new set of worklists for garments in museum collections. - PKM (talk) 22:44, 23 October 2020 (UTC)[reply]