Property talk:P4663

Documentation

DACS ID (former)
obsolete code used to identify 50,000 artists as members of the British collective rights management organisation DACS and sister organisations worldwide

Associated item

Design and Artists Copyright Society (Q3705963), Design and Artists Copyright Society (Q3705963)

Applicable "stated in" value

DACS register (Q71534274)

Data type

External identifier

Domain

artists (note: this should be moved to the property statements)

Allowed values

[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}

Example

Albert Dandoy (Q24654988) → 3bf1d251-51e5-e711-8b59-000c29e811b2
Andy Warhol (Q5603) → 6857c30f-cb65-4950-84ef-bb52b22927b5
Frida Kahlo (Q5588) → 732fb75a-7761-4541-9df7-b9acf5335109

Source

https://www.dacs.org.uk/licensing-works/artist-search

Formatter URL

https://web.archive.org/web/20210630075411/https://www.dacs.org.uk/licensing-works/artist-search/artist-details?ArtistId=$1

Formatter URN

urn:uuid:$1 — List of Uniform Resource Names (URN)

Tracking: usage

Category:Pages using Wikidata property P4663 (Q45892284)

Related to country

United Kingdom (Q145) (See 324 others)

Those IDs do not work edit

Latest comment: 6 years ago16 comments4 people in discussion

@Hannolans, Jheald, Multichill, ديفيد عادل وهبة خليل 2, Pintoch, Magnus_Manske:

Hello,

while trying to complete Mix'n'Match for some artists, I stumbled on this ID... but clearly : either they changed the site, either the adress is wrong... all pages are blanck and do not give access to artist's pages :(

Maybe this catalog should be de-activated in Mix'n'Match, untill the problem is solved ? --Hsarrazin (talk) 18:18, 21 January 2018 (UTC)Reply

Hmm, I checked and Andy Warhol, the example is not working anymore.... We hesitated to create this property as this had been altered some years ago. We monitored some months and it seemed to stay persistent. Needs further research..--Hannolans (talk) 19:02, 21 January 2018 (UTC)Reply

They have other identifiers now,Andy warhol ID is now: 6857c30f-cb65-4950-84ef-bb52b22927b5 --Hannolans (talk) 19:05, 21 January 2018 (UTC)Reply

@Hannolans, Hsarrazin: Yes. Like you, I was doing some Mix'n'match a couple of days ago, and found exactly this. It ought to be reasonably straightforward to do a re-scrape; but I don't know if it might or might not then perhaps be possible to match the names and text to the previous names and text. I did post to Magnus about it; but of course he's very busy -- and we need to wonder whether, given that I think this is already the second time this has happened, will the IDs be any more stable this next time. It may be worth emailing DACS to ask. Jheald (talk) 01:08, 22 January 2018 (UTC)Reply

@Jheald: Yes. I agree that maybe it's not worth it. Asking DACS would be good, to know whether they think it might be interesting to keep stable IDs, and then be referenced on wikidata, or if they don't really care. Their Terms and conditions do not make me very optimistic. --Hsarrazin (talk) 10:19, 22 January 2018 (UTC)Reply

I have sent an email to DACS, no response yet. As this data of DACS is very useful to link to the copyright status of authors. --Hannolans (talk) 20:34, 1 February 2018 (UTC)Reply

Me too. I wrote a couple of days ago, but nothing back yet. If necessary, it should be possible to scrape the new site, and match names & descriptions with what was loaded into Mix'n'Match from the last scrape. Jheald (talk) 20:38, 1 February 2018 (UTC)Reply

@Hannolans: Just had a reply back from their Digital Projects Manager:

Thank you for your enquiry and interest in DACS.

As a dynamic organisation we regularly review our website and make changes, and cannot guarantee the consistency or stability of links. We do endeavour to keep the information up-to-date, but frequent modifications or deletions of entries will continue to be a regular occurrence.

So I guess we're on our own. :-)

I'll see what I can do to make an old->new mapping, based on the old information in MnM.

Going forward, it's probably a sign that the site will need to be routinely re-scraped occasionally, to validate that old links still work, and see if any new ones have popped up. Not such a big deal, we're in a similar position with eg Art UK, where the identifiers change eg if they find out that an artist has died. I'll try and progress this over the next couple of days. Jheald (talk) 16:23, 2 February 2018 (UTC)Reply

@Hannolans, Hsarrazin: Updated values now going in with QuickStatements for the 2000 items we have that are matched; and I've sent Magnus some files to update QuickStatements. That leaves about 187 which we did have matched, but that I wasn't able to automatically identify the new values for, in most cases I think because DACS have modified the spelling of the name. The easiest thing, I think, is just to go through these manually and look up the new values using the search function on the DACS website. I'll see if I can get that done tomorrow. Jheald (talk) 00:36, 10 February 2018 (UTC)Reply

Great work! I altered the constraints regex and did some clean up. --Hannolans (talk) 21:38, 11 February 2018 (UTC)Reply

Reminder to remove the mandatory constraint status until the cleanup is done, otherwise various reports will be unusable. Sjoerd de Bruin (talk) 16:37, 13 February 2018 (UTC)Reply

Data is all clean, I am waiting for a new constraint report if we missed something. --Hannolans (talk) 08:05, 14 February 2018 (UTC)Reply

@Hannolans, Sjoerddebruin: No, the data is not clean; though it passes the constraint check. There are still currently 154 items with the old-style DACS identifiers:

SELECT ?item ?itemLabel ?dacs WHERE {
  ?item wdt:P4663 ?dacs .
  FILTER (!regex(?dacs, '-')) . 
  
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

Try it!

I was waiting for Magnus to load the new values up in Mix'n'Match, and then to go through with MnM search to match them. But I'm not sure if he saw the links I posted on his talk page to the relevant files to update the values at MnM without losing existing matches. Jheald (talk) 11:27, 14 February 2018 (UTC)Reply

@Hannolans, Hsarrazin, Sjoerddebruin: Magnus has worked lots of magic, and the new IDs are now up as catalog #1001 on Mix'n'Match. The old matches have been carried forward, including their attributions, except for 75 (currently) that were not possible to map forward from old to new automatically, usually because the names have been changed too much.

These can be found in a Listeria table at Property talk:P4663/old ids, from which they will disappear as the old IDs are removed. It should be possible to find new IDs, either by searching at DACS, or on Mix'n'Match. In MnM search, the relevant catalogue is "DACS (Feb 2018)", currently appearing right at the bottom of a name search. There may also be another catalogue below it, "DACS (2018)" -- this should be ignored, it was an error that I made when uploading. The old IDs can be safely removed once new ones have been found. With luck the names should be reasonably close, but if not it would be worth trying usual MnM tricks, such as removing some names, and shortening others that are left, by removing the end and replacing it with *. MnM search can be quite slow to load, so I find it can also save time to open about 5 MnM search links at a time, in tabs. Thanks to everybody for their patience, best wishes Jheald (talk) 19:15, 19 February 2018 (UTC).Reply

This query tinyurl.com/ybejytve can be used to find items with both statements, eg as a step to removing the old one using Quick Statements 2. Jheald (talk) 17:01, 21 February 2018 (UTC)Reply

Update. 3 names now remaining with the old IDs, each a bit difficult to search for. Jheald (talk) 13:10, 2 March 2018 (UTC)Reply

Offline edit

Latest comment: 4 years ago1 comment1 person in discussion

Note that today DACS author seems offline, see for example https://www.dacs.org.uk/licensing-works/browse-artists?Letter=A --Hannolans (talk) 18:07, 19 December 2019 (UTC)Reply

correct P31 value of entries edit

Latest comment: 3 years ago1 comment1 person in discussion

This finds some 240 entries with instance of (P31)=human (Q5) and a label that suggests the items is not about a person (or only amalgamated with a person). I came across them when trying to add given name (P735) to these items. @Hannolans: --- Jura 13:21, 14 January 2021 (UTC)Reply

Form of identifier has changed edit

Latest comment: 4 months ago3 comments3 people in discussion

The form of the identifier for this property has changed again. Instead of [0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12} the current form is ARTS\d{6}. I corrected the examples and the corresponding items, but all of the other items with this property will need to be fixed as well. UWashPrincipalCataloger (talk) 16:54, 14 March 2022 (UTC)Reply

Hmpf. We have 65k statements... Could we update Mix n Match? If we have a scraped dataset I can do some mass edits. Hannolans (talk) 20:06, 15 March 2022 (UTC)Reply

New property DACS ID (2022) (P10706). Most previous matches now mapped forward. Jheald (talk) 11:41, 16 December 2023 (UTC)Reply

Removal request for BLP date of birth statements sourced to old version of DACS register edit

Latest comment: 3 months ago13 comments4 people in discussion

I've had a request from DACS to take down all date of birth (P569) statements that are sourced to the old (P4663) version of the DACS register.

DACS no longer disclose dates of birth of living people on their website. However, we are still showing dates of birth for about 14,000 of their members, single-sourced to DACS register (Q71534274) and a DACS ID (former) (P4663) value. It would seem that some of their members are not happy about this, and have told DACS to make it disappear; so DACS has asked for all P569 statements sourced from their old pages to be removed.

I pinged the Wikidata telegram channel for thoughts, and the quick take there was that DACS has raised a serious issue: per our WD:BLP policy, values of 'sensitive' information for living people (including years of birth) "should generally not be supplied unless they can be considered widespread public knowledge or are openly supplied by the individual themselves". While arguably this may have been satisfied when the information was readily being made publicly available as part of the DACS register, the view was that this could no longer suffice for WD:BLP if the dob information was no longer being made readily available.

I have suppressed the dates of birth for ten Irish artists https://w.wiki/8W4e that DACS had received specific objections about.

They have accepted that it is reasonable for us to see whether we can find alternative sourcing for the remaining 14,000 DoBs currently single-sourced to P4663, which will need a bit of time; but ask to be kept advised of our progress. These 14,000 can be found with query https://w.wiki/8UTN (there may be a few more in addition, that are sourced to P4663 more than once).

ADDED: Ooops. Looks like that adds another ~16,000 : https://w.wiki/8W6W

I've currently also got a QS run going through to add

on focus list of Wikimedia project (P5008)DACS BLP 2023 (Q123824574)

to make these easier to recall and analyse.

With luck we should be able to find alternate sourcing for at least some of these DoBs (VIAF, ULAN, refs in Wikipedia articles, artist websites, etc...); but it's going to take a bit of searching. Jheald (talk) 23:08, 14 December 2023 (UTC)Reply

Some queries:

https://w.wiki/8WWh -- Items tagged for the focus list. -- Baseline (2023-12-15): N = 29,696 (with names & dobs: https://w.wiki/8eqn) -- items below Q96000000 possibly should not be in list
https://w.wiki/8W5x -- Number of IDs for items in the target set. -- Baseline: 27,895 (93.9%) have no other ID.
https://w.wiki/8W5y -- Most common other IDs
https://w.wiki/8W62 -- Number of sitelinks for items in the target set. -- Baseline: 29,225 (98.4%) have no sitelinks.
https://w.wiki/8W66 -- Most common sites for sitelinks
https://w.wiki/8WMa -- Count of items with no sitelinks and no other IDs. -- Baseline: 27,762 (93.5%) have no other ID or sitelink.
https://w.wiki/8WPA -- Number missing either a family name (P734) or given name (P735) coding. -- Baseline 7,117 (24.0%)
https://w.wiki/8WQS -- Scan for possible matches based on matching P734 & P735 names + dobs within 500 days (needs to be stepped through set) -- looked from offset=12000 (and higher)
https://w.wiki/8WQk -- Similar, but with Soundex matching of surname
https://w.wiki/8WX7 -- Tagged items with a different non-DACS date of birth -- Baseline: 96 (alt dobs either conflict or are unreferenced)
Most common countries https://w.wiki/8e5f ; years of birth https://w.wiki/8e5p

Jheald (talk) 23:31, 14 December 2023 (UTC)Reply

@Hannolans, Multichill, Hsarrazin, Edward, Vojtěch Dostál: Pinging you in case you have any good ideas or tools to try to locate some alternative sources for these DoBs.

It would be good to harvest any dobs from any existing matched IDs or sitelinks; but, per the stats above, only rather few of the target items have such matches (and perhaps even fewer when we clear up the set with alt dobs that are unreferenced).

So a priority may be if anyone has any good tools to try to identify some external matches. Or any duplicates inside wikidata -- the query based on similar family names + given names + dob suggests there might be some; and perhaps it could be generalised ?

Anyone got any thoughts, or able to help ? Jheald (talk) 15:32, 15 December 2023 (UTC)Reply

The Fine Art Archive (Q107456632) may have quite a few of them. Skim was most recently in touch with them - Michal, do we have a copy of their database so that we can look if some of these DoB could be alternatively sourced from there? Vojtěch Dostál (talk) 19:17, 15 December 2023 (UTC)Reply

Also, the query above should filter out people which do have another DoB statement and with a non-DACS reference, such as Alena Janoušková Vančurová (Q99477823). Vojtěch Dostál (talk) 19:20, 15 December 2023 (UTC)Reply

@Vojtěch Dostál: The latter group should be excluded from the DACS BLP 2023 (Q123824574) tagging (eg the item for Vančurová doesn't have the tag), apart from cases where the second DoB statement is not referenced; or is only year-precison and doesn't agree about the year. (See https://w.wiki/8WX7 for these, currently about 100 to examine). Jheald (talk) 22:01, 15 December 2023 (UTC)Reply

@Jheald I see. Yeah, unfortunately most of these items don't have many useful identifiers which could be used as an alternative source :( Vojtěch Dostál (talk) 09:06, 17 December 2023 (UTC)Reply

@Vojtěch Dostál: Not yet, anyway. Not unless we can match them. Jheald (talk) 10:03, 17 December 2023 (UTC)Reply

Update: https://w.wiki/8fZm 16,403 birth dates now second-sourced to ADAGP. That leaves about 13,200 to go. (https://w.wiki/8id6 : 13,828; https://w.wiki/8idC 15,757; diff https://w.wiki/8idU due to dates of death -- or almost https://w.wiki/8jA6) Jheald (talk) 22:42, 29 December 2023 (UTC)Reply
Who gives you the right to start doxxing people by collecting their age and publishing it without their consent?

This is personal data that you have no right to be either accessing or publishing. Have the individuals given you consent under EU GDPR for you to process their personal information? I suggest you should seek legal advice.

I have taken note of the individual names above and formal complaints been submitted regarding their actions to both UK and French Data Protection Authorities. 80.233.72.63 22:39, 3 January 2024 (UTC)Reply

@80.233.72.63: Thanks for getting in touch. Direct communication is always helpful, and I am sorry that you were unhappy to find your date of birth on Wikidata.

I remain of the view that a year of birth is in general a valuable thing to include in a biographical index of artists. As well as giving a general sense of the periods that somebody may be associated with, for a project like wikidata which exists to try to connect different sources to each other it can be particularly valuable, for helping to determine when two different sources may be referring to the same person; and equally to help identify and prevent links about different people becoming conflated in the same record, which is in nobody's interest.

This is why most similar indexing projects try to collect and add date of birth information where publicly available -- for example, library catalogues such as the Library of Congress (eg [1]), VIAF [2], ISNI [3], the Bibliothèque nationale de France [4] ; art history indexes such as ULAN [5], RKD [6], Benezit [7], the NYPL index of photographers [8]; and collecting societies such as ADAGP [9] all try to give at least a year of birth where possible.

For this reason, I believe that including this information does serve a "legitimate interest" in the language of the GDPR legislation, and that continuing to hold it does serve a valuable ongoing useful function.

The GDPR does not in general require pre-authorised consent [10]; though information about living people should indeed be treated with special care, as the wikidata community recognises (see the Wikidata:Living people policy, as was also linked above).

In this case, the year-of-birth information (originally sourced from DACS) for the ten individuals notified to me by DACS as being of particular concern was down the very same evening despite it being a handful of days before Christmas -- despite in 5 of the cases the very same information being readily available in gallery exhibits, online CVs, or books, presumably as authorised by the artist; and in one case even being given on the 'about' page of the artist's own website.

I also added a flag for the absence on the page (eg [11]) to try to stop anyone else innocently re-adding the information, with a reference that it seems was successful in helping you find this coordination page.

I am very open to any concerns or suggestions you may have about the Living people policy, which like all wiki policies is a living document, so open to improvement; and to any suggestions you may have about the information we have on individual artists' pages and how it is presented (though I am only one contributor, so my ability to change things is limited).

For better or worse though I think it needs to be recognised that year of birth has become the standard first-line way to distinguish an artist from another artist or another person with a similar name, so I do think it is appropriate to try to include here, where it has been publicly presented, while being very ready to suppress it in cases where that is requested. Jheald (talk) 18:51, 4 January 2024 (UTC)Reply

Hi !

I would gladly help in sourcing dob from other sources, and adding other IDs in the same move (like on Pierre-Paul Hueber (Q99456028) or Odette Allouard-Carny (Q99461858), but my sources are specialized in French artists.

1. Could you please provide a list of French artists only ?

2. when another source than DACS is found, do I need to remove DACS source ? or statement when less precise (year/full date) ?

Thanks in advance :) Hsarrazin (talk) 16:32, 8 January 2024 (UTC)Reply

@Hsarrazin: Thank you so much. Here's a current list of artists with country of citizenship (P27) = France (Q142) and dates of birth referenced only to DACS: https://w.wiki/8p5o Please leave the DACS BLP 2023 (Q123824574) tag and the DACS source for the moment while work on the set as a whole is going forward -- it helps confirm that the old DACS ID was correctly matched, and is helpful for progress tracking; also the DACS source is less objectionable when other sources are also present -- though in time it is likely the DACS source will likely be removed, once as much clean-up work as we can do (including checking any unlikely-looking old ID matches) has been done across the whole dataset. Jheald (talk) 17:01, 11 January 2024 (UTC)Reply

possible errors found edit

I begin here a list of items probably created from erroneous records : typos or other...

Talk:Q99460661 is probably Maurice Estève (Q471318)

Add topic