Wikidata talk:WikiProject Archival Description

ISAD(G) or RiC?

Hi, I would rather suggest to go with the approach taken by Records in Context (RiC), which has the ambition to make the data models for archival description ready for the linked data world.

Concretely, the sole classes above the item would be:

• collection
• record set

They would be related among each other through <part of>/<has part> relationships.

Furthermore, a new property could be created to indicate the <level of description> (according to ISAD(G)). Before doing that, I would however ask to what extent the ISAD(G) levels of description indicated in archival finding aids share some common semantics across institutions (or conversely: to what extent does each archive have its own idiosyncratic approach to applying levels of description?). - If there is no shared sematics with clear definitions for the different levels of description, it may not make much sense to ingest that data into Wikidata.

Cheers, Beat Estermann (talk) 06:24, 24 April 2018 (UTC)

See in this context also: Matterhorn RDF. --Beat Estermann (talk) 08:47, 26 September 2018 (UTC)

Current usage of archives at (P485)

# List pairs of items / archives linked by the property P485 (archives at)
SELECT ?item ?itemLabel ?typeLabel ?archives ?archivesLabel ?archivestypeLabel
WHERE
{
  ?item wdt:P485 ?archives.
  ?item wdt:P31 ?itemtype.
  ?archives wdt:P31 ?archivestype.
     
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en". }
}

Try it!

Studies regarding the inventory building in the area of audiovisual heritage

The following two studies should be considered when discussing how to describe record sets in Wikidata:

• Schade, E., Aschwanden M. (2017) Evaluation der Inventarisierung audiovisueller Kulturgüter in der Schweiz durch MEMORIAV (Lagebericht als Teilstudie 1), HTW Chur.
• Schade E., Mastrandrea, E. (2017) Evaluation der Inventarisierung audiovisueller Kulturgüter in der Schweiz durch MEMORIAV (Visionen als Teilstudie 2), HTW Chur.

The second study contains a series of metadata sets that are described from the perspective of their functional requirements. I'll see whether I can get it published.

--Beat Estermann (talk) 07:02, 31 May 2018 (UTC)

infoclio.ch database

The infoclio.ch database could be an interesting starting point. It is expected to be released under an open license in view of the next #GLAMhack. --Beat Estermann (talk) 06:17, 7 June 2018 (UTC)

level of description (P6224)

Hi all,

There is now a property for level of description (P6224), which can be used to better describe archival holdings. Before going out and using it on a bunch of pages, I think it would be useful for us to discuss how it should be used. I think it would be best if we identify Wikidata items for all possible levels of description by adding "instance of (P31)":"level of description (Q59211454)" to those, so we have a list of the options for level of description. I went ahead and added 5 basic terms (see here). Are there others on Wikidata already, or other suggestions? Dominic (talk) 23:35, 10 December 2018 (UTC)

Thanks for intervention. At this moment some levels of the "level of description" we have here (and it is advisable to merge them). For level "item": item (Q11723795), for level "series": series (Q3511132), for level "collection/fond": fonds (Q3052382) or archival collection (Q9388534). Maybe more ambiguious wikidata item: act (Q421829) (for level "file unit/file"). When You are making order in levels of description, can you check this issue and merge them (at least first two low level are clear)? --marv1N (talk) 20:49, 11 December 2018 (UTC)

@Dominic:: Good news, I made the sugested clearing. So, the main levels are record group (Q59294700) > fonds (Q59294612) > series (Q3511132) > file unit (Q59221146) > item (Q11723795) (at the end, I merged only Q3511132 and Q11723795, rest wasn't exactly about categories of the levels of desrcription, so I let them be...).  – The preceding unsigned comment was added by [[User:|?]] ([[User talk:|talk]] • contribs).

• Please see also Property talk:P1436 about the number of elements that have been described. --- Jura 14:29, 26 August 2019 (UTC)

Page Typology

Hi ! I am a new contributor on Wikidata and followed a course. As an archivist, I wish to contribute to the project. My first contribution: creating the Typology pages witch wishes to gather the diffrent types of archives in Wikidata. Feel free to add query and things ! --LuciOle (talk) 18:06, 12 August 2019 (UTC)

To typology of archives, we have start here: Wikidata:WikiProject_Heritage_institutions/Typology#Proposed_thesaurus_for_archives. --marv1N (talk) 18:29, 12 August 2019 (UTC)

I saw it and I took some informations from it. But I also meant we could maybe think about typology of archives in terms of private archives, public archives, archives with audiovisual material, press archives, etc. Something to reflect on ! --LuciOle (talk) 16:46, 13 August 2019 (UTC)

I think, that it is good idea to have "archival topics" here (or there) together - on the other hand this typology here should be more or less the same as on WikiProject_Heritage_institutions (for example private archive (Q12161242), press archive (Q56650887) is omitted only by mistake?, not on purpose). --marv1N (talk) 19:02, 13 August 2019 (UTC)

Yes, some are missing. I will try to add them. --LuciOle (talk) 10:25, 3 September 2019 (UTC)

Scope and content

See: Wikidata:Property proposal/scope and content

As the character limit for monolingual text properties is 1500 characters (essentially half a page of text), I have decided to propose a new property for "scope and content," since this is such an important element in archival description. I would like to invite you all to check out the property proposal. Once created, I would plan to import all scope and content notes for NARA archival descriptions into Wikidata. Dominic (talk) 13:21, 30 August 2019 (UTC)

Modeling dates

Weadock313 (talk) 01:27, 20 April 2020 (UTC) 2le2im-bdc (talk) 19:36, 30 May 2018 (UTC) Beat Estermann (talk) 23:29, 1 June 2018 (UTC) Flor WMCH Gilliane Kern (talk) 05:47, 5 June 2018 (UTC) Mauricio V. Genta (talk) 23:42, 1 September 2018 (UTC) Laureano Macedo (talk) 09:49, 13 September 2018 (UTC) Daniel Mietchen VIGNERON Patafisik anarchivist KelliBee123 (talk) 16:29, 6 August 2019 (UTC) :kedouch Kedouch (talk) 11:43, 6 December 2023 (UTC) Anne Chardonnens (talk) Yooylee 30 (talk) 23:15, 11 August 2019 (UTC) Mlemusrojas (talk) 19:40, 19 February 2019 (UTC) erussey Kcohenp (talk) 16:28, 03 June 2019 (UTC) Mrtngrsbch Amandine (talk) 19:58, 12 August 2019 (UTC) RenéLC (talk) 11:17, 13 February 2020 (UTC) Sp!ros (talk) 09:00, 27 April 2020 (UTC) Ccooneycuny (talk) 16:49, 19 May 2020 (UTC) Librarian lena (talk) 19:57, 11 August 2020 (UTC) Valeriummaximum (talk) 16:42, 23 August 2020 (UTC) Jneubert (talk) 07:11, 24 September 2020 (UTC) MaryCDominique (talk) 10:06, 25 October 2020 (UTC) Epìdosis 17:16, 16 February 2021 (UTC) P feliciati (talk) 17:18, 16 February 2021 (UTC) JnyBn (talk) 07:27, 22 April 2021 (UTC) Hsarrazin (talk) 13:25, 1 July 2021 (UTC) Carlobia (talk) 15:27, 6 October 2021 (UTC) Heberlei (talk) 20:19, 29 July 2021 (UTC) KAMEDA, Akihiro (talk) 18:30, 18 September 2021 (UTC) Jonathan Groß (talk) 08:41, 1 November 2023 (UTC)

Notified participants of WikiProject Archival Description

NARA's data standard contains the following guidance about the two types of ranged date fields for collections:

Inclusive Dates Versus Coverage Dates

• Inclusive dates are those that comprise the time period during which the record group, collection, or series was created, maintained, or accumulated as a unified filing system by the creator because of some relationship arising out of the records creation, receipt, or use.
• Coverage dates are those that comprise the subject time period covered by the record group, collection, or archival materials. Normally, the subject dates are the same as the inclusive dates. However, if the dates are different, use the coverage dates as appropriate.

For "coverage dates", we have start of covered period (P7103) and end of covered period (P7104), which fits the meaning nicely. But for the "inclusive dates", this seems like a property that is more specific to the archival context, and I cannot find a good fit. Does anyone have any suggestions for how to model this? Dominic (talk) 13:43, 30 August 2019 (UTC)

Place access points

Hi all !

Investigating a migration of archival data to Wikidata, I don't find a suitable property for Place Access Points. Would any of you know of a property that correspond ? So far, the best I have found is [(depicts)], but it is not totally satisfying. I am searching for a property that would be something like "Mentions location"...any help would be greatly appreciated !

Vdelavenne (talk) 20:22, 3 October 2019 (UTC)

@Vdelavenne: You can use depicts (P180) on archival works, even non-graphical ones, like a textual document. You can also use main subject (P921) is appropriate, though not all mentions in a work are a primary subject of a work. Otherwise, there is a property inscription mentions (P6568) that seems closer to what you are looking for (you could put in a location, if relevant), but it is for inscriptions on plaques/monuments only, and I don't see a similar property for other types of works. Might be useful to have a discussion about this, either generalizing that property's scope or proposing a new property. Dominic (talk) 15:08, 7 October 2019 (UTC)

How to model multiple objects in single inventory item

Hello! I am working with a large collection of data for archaeological items and am facing a very specific issue of how to deal with items that sometimes contain multiple objects (i.e. a collection of rings or amulets or fragments of a painting that have been inventoried under the same number). In the metadata record, in cases where multiple objects are represented as the same item, we have multiple values i.e.

dimensions (height x width): a: 5x5cm, b: 7x4cm, c: 16x4cm 

I am just wondering how this should be modelled in wikidata? Should each object be a separate item in Wikidata (even if designated by the institution under the same inventory number), or should multiple values be entered for height (i.e. height: 5, 7, 16)--though this would mean losing the index relation (i.e. object a has height 5cm) Perhaps a relevant concern, I am not sure if objects that exist under the same inventory number are necessarily related to one another (sometimes archivists and antiquarians just lump similar things together in the same box...) Would love to hear thoughts on best practices! Valeriummaximum (talk) 18:53, 23 August 2020 (UTC)

• Also, on a related note, how would you model an item for which the dimensions distinguish the frame of the object, the size of a figure on it, or any other smaller part of the object?Valeriummaximum (talk) 19:08, 23 August 2020 (UTC)

@Valeriummaximum: Ideally, each part would have its own item (I often see the inventory numbers 1234a, 1234b, or 1234.1, 1234.2). If that doesn't make sense in a particular context, you can use "applies to part" to qualify the various sizes, dates, materials, etc. - PKM (talk) 22:10, 29 November 2020 (UTC)

• • thank you PKM. In the end, I went with 'applies to part' and specified 'item' if it applied to a component item (or, similarly, 'vase','bottle',etc, and 'handle','neck','maximum', if those specifications existed in the data). Unfortunately not every item has component IDs for its subcomponents and there is also the danger that some items may be split and be given separate gallery ID numbers, so it will be an ongoing challenge to maintain these records and update them in Wikidata. Valeriummaximum (talk) 13:39, 30 November 2020 (UTC)

scope and content (P7535)

Somehow the free text experiment with the above property doesn't quite work out. samples

• Q64853740 "Plan et élévation de l'Autel de la Sainte Vierge projeté pour la cathédrale de Limoges"
• Q64853741 "Cathédrale de Limoges": plan général.

Despite assurances that "trained archivists at the repositories" could handle it (see Wikidata:Property proposal/scope and content), we get values like the above, mainly duplicating information already available in a structured way. These amount to almost all uses of the property. Unless this can be remediated in one way or the other, I'd propose the property for deletion.

@Archives nationales DJI, Dominic, ArthurPSmith, PKM, SCIdude: @Anarchivist, Nomen ad hoc, 2le2im-bdc, Wskent: --- Jura 13:29, 29 November 2020 (UTC)

• What is the issue here? These appear to be separate plans for the cathedral with distinct inventory numbers Valeriummaximum (talk) 13:44, 30 November 2020 (UTC)

There is no added value in the repetition of item or P170 labels.

I don't see how it meets "A narrative statement summarizing the characteristics of the described materials, the functions and activities that produced them, and the types of information contained therein.". Even the items' description is closer. --- Jura 14:49, 30 November 2020 (UTC)

I'm not sure I understand still--what item is repeated?Valeriummaximum (talk) 15:01, 30 November 2020 (UTC)

sorry if I didn't understand above--the issue is about the string text for P7535 not conveying unique information? it seems here that P7535 shouldn't be used here to begin with and wasn't intended as a labelValeriummaximum (talk) 16:50, 30 November 2020 (UTC)

Use of Q3052382 (archival fonds)

Hi, I see that there is a value "fonds" yet, it doesn't seem widely used compared to the property "archives at" to identify archives. While this property "archives at" is of great use it seems to me that creating new items for the different archival fonds of my institution would also be useful. I created one of such items here. I'd love to hear other archivists opinion about the creation of such items. ping : User:2le2im-bdc, User:Gilliane, User:Cosovschi--RenéLC (talk) 14:29, 29 September 2021 (UTC)

Archival fonds and the institutions owning them

Considering the items regarding archival fonds, i.e. having instance of (P31)fonds (Q3052382), they are owned by institutions, usually archival institutions. As of now I think there are at least 4 different, and conflicting, ways in which we model this relationship:

1. 10741 cases of fondscollection (P195)archive (https://w.wiki/7hgt)
2. 9809 cases of fondslocation (P276)archive (https://w.wiki/7hgr)
3. 10030 cases of fondsarchives at (P485)archive (https://w.wiki/7hgv)
4. 1226 cases of fondscollection (P195)archive + fondslocation (P276)archive (https://w.wiki/7hh8)
5. 0 cases of fondslocation (P276)archive + fondsarchives at (P485)archive (https://w.wiki/7hh2)
6. 0 cases of fondscollection (P195)archive + fondsarchives at (P485)archive (https://w.wiki/7hh6)

My proposal is adopting univocally option 1, suppressing options 2, 3, 4 in favour of it. Do you agree? --Epìdosis 18:28, 25 October 2023 (UTC) FYI

Angayubagia CRolker Epìdosis HHill Jahl de Vautban Jason Evans Jonathan Groß Lynn Ransom Marsupium MartinPoulter Mel22 Pmt PKM Toby Burrows Maxime

Notified participants of WikiProject Manuscripts

I agree that option 1 is the best fit. location (P276) should only be used if the archive has more than one location (subsidiaries, reading rooms in other buildings) and we have a specific value (i.e. a dedicated item) for this location. Jonathan Groß (talk) 18:32, 25 October 2023 (UTC)

I'm now suppressing option 4 (the most redundant among all) through https://quickstatements.toolforge.org/#/batch/215365. I wait tomorrow for the fix of options 2 and 3. --Epìdosis 06:40, 27 October 2023 (UTC)

I started a partial fix for option 3 (the cases without reference, so the easiest ones) through https://quickstatements.toolforge.org/#/batch/215443. --Epìdosis 16:03, 27 October 2023 (UTC)

Now fixing half of the remaining option 2 through https://quickstatements.toolforge.org/#/batch/215607. --Epìdosis 16:12, 29 October 2023 (UTC)

Option 3 will be completely fixed through https://quickstatements.toolforge.org/#/batch/215610. --Epìdosis 16:23, 29 October 2023 (UTC)

Option 2 will be finally fixed through https://quickstatements.toolforge.org/#/batch/215660. This completes the transition. Good night, --Epìdosis 23:58, 30 October 2023 (UTC)

I agree with preferring option 1 and correcting the others. - PKM (talk) 23:53, 29 October 2023 (UTC)

Let's set a few things straight before we move forward...

As part of my quest for the best place to document the best practices regarding the use of the "archives at", "documentation files at", etc. properties (based on this Google Doc), I checked out the project "Archival Description" as well as some of the data structures it is supposed to deal with (see my notes in this Google Doc). Thereby, I encountered the following inconsistencies and open questions that should be addressed before we moved forward:

• The name of the WikiProject is not in line with its scope: "The aim of the present project is to create the world’s most comprehensive high quality database of archival fonds and heritage collections,..." - I would advocate in favor of sticking to the broader scope, englobing all heritage collections; the name of the project should be adapted accordingly.
• There is confusion as to the definition of some of the core concepts of this project. Most importantly, as the previous discussion item demonstrates, there is total confusion as to the defintion of an "(archival) fonds". fonds (Q3052382) is currently defined as a "level of description"; at the same time it is widely being used to describe "record sets" (I am using the term in the sense of Records in Context RiC). - I would advocate in favor of using an approach following Records in Context: At the top level, we would have collection (Q2668072) and archives (Q56648173); at the bottom level, we would have individual items, and anything in between would be a "record set", which conceptually would be a subclass of unit of description (Q117451989). "Record sets" could have a "level of description" attribute that follows the definition of ISAD(G) and its adaptations. Some adaptations (e.g. the Austrian one) foresee a supplementary level of description above the "fonds"; this shouldn't be a problem. However, I have no idea how best to deal with the fact that currently, hundreds of thousands "collection" items have a level of description referring to the archival world. I would suspect that most of them are not in line with the current definition of collection (Q2668072) as a "set of purposefully gathered physical or digital objects with some common characteristics"; to me, this definition refers to the principle of "pertinence", while for the archival record sets, I would assume that the principle of "provenance" is much more widespread.
• The relevant class trees as well as the definitions (including the <instance of> statements> for heritage institutions, and collections/archives currently seem to be a complete mess: a "library" (institution) is defined as a subclass of a "collection"; a "collection" is defined as an instance of <subject heading>, <academic discipline>, <speciality>, <field of study>, etc. - They need to be tidied up; relevant items should be "claimed" (maintained) by this WikiProject.
• When it comes to describing heritage holdings, we suggest two rather different approaches: (1) full description on Wikidata; (2) external reference via "archives at", "documentation files at", etc. properties (see: Wikidata:WikiProject Performing arts/Typologies#Artefacts Documenting Activities Related to the Performing Arts for what is currently probably the most comprehensive description of this approach). We do not explain when to use which approach, and I so far have not found any proposal how best to combine the two approaches (should they be exclusive or complementary for a given record set?) - The best practice in this regard should be clarified. I do not think that approach number 1 scales well. In my view, it should be used whenever a given record set also has (or is supposed to have) an entry on Wikipedia. In all the other cases, I would suggest that we opt for approach number 2 exclusively. Views may differ on this, and it does not hurt much if someone does not follow this recommendation. More importantly, we should come up with a best practice as to how the two approaches should be integrated or used in a complementary fashion: Someone searching for heritage holdings on Wikidata should ideally find all of them. The ideal solution would be to find a way to make sure that all the holdings described according to approach number 1 also show up in queries following the approach number 2. - Are there any thoughts on how best to achieve this?

--Beat Estermann (talk) 12:28, 27 October 2023 (UTC)

I am reluctant (in general) to restrict item creation in Wikidata based on article creation in Wikipedia. I do agree that "archives at" should be used on people and organizations. But I think archives that meet WD notability criteria can (should?) have their own items. PKM (talk) 00:14, 30 October 2023 (UTC)

@PKM: I think this is not what Beat Estermann suggested. He wants to restrict the description of "heritage holdings", not creation of items for holding institutions. In fact, if we follow Beat's recommendation (which I think I could get on board with), we will need to create quality items for holding institutions to have a target value for archives at (P485). Jonathan Groß (talk) 08:59, 1 November 2023 (UTC)

Ah, my mistake. I am all in favor of adding new items for archival institutions! PKM (talk) 22:47, 1 November 2023 (UTC)

@Beat Estermann: Thanks for starting this discussion and your clear argument. I'm afraid I can't contribute to all points raised, as I am no archivist, but I will give you my opinion on some aspects as I hope it helps.

• Regarding the name-scope-discrepancy: When I found this project I thought its goal was to align Wikidata Data Modelling with state-of-the-art archival description. It would be great if we had a dedicated project for that, and part of your argument goes in this direction if I am not mistaken. Adopting a more general name like Wikidata:WikiProject Archives is  a good idea.
• The confusion of terms is something that should be addressed sooner rather than later, and ideally by an archivist (or a team of archivist)-Wikidata user(s) with good knowledge of the standards of the field across different countries. I am sure there are many competing standards, but I also image there must be some common ground which can be adapted to our data model. Maybe there is even an internationally recognised Minimalkonsens we can use?
• Cleaning up the class tree should be its own sub-project within this WikiProject.
• I  support your approach number 2 to describing heritage holdings, provided that it does not in principle prohibit creating items for (notable) heritage items or record sets and connecting them with pertinent items in a sustainable way. Best, Jonathan Groß (talk) 10:02, 2 November 2023 (UTC)

Regarding the name-scope-discrepancy: We already do have a Wikidata:WikiProject Heritage institutions; I therefore don't think creating a Wikidata:WikiProject Archives would make much sense. At the same time, I believe that it would be useful and most straightforward to document approach number 2 in a way that covers all types of heritage (and scientific) collections (not just archival holdings) in one place. I was briefly considering extending the scope of Wikidata:WikiProject Heritage institutions accordingly; but that might just overload that other project. Maybe it would be best to create a new Wikidata:WikiProject Heritage Collections to document approach number 2. This project (Archival Description) could keep its name, and we could restrict its scope to issues related to the full description of archival holdings following approach number 1. -- Beat Estermann (talk) 07:27, 16 November 2023 (UTC)

I have started to set up a new WikiProject Wikidata:WikiProject Heritage Collections, with the intent of moving the entire documentation regarding the use of “archives at”, “documentation files at”, etc. to this new WikiProject and eventually removing it from Wikidata:WikiProject Archival Description, making it possible to dedicate the latter to the full description of archival holdings on Wikidata (approach number 1). -- Beat Estermann (talk) 13:15, 21 December 2023 (UTC)