MIDS Element - 2. Organization

[RBGE-Herbarium]

Information Element Name	Organization
Modified	08/12/2023
Label	Organization
Definition	A term to indicate in which institution the specimen is held. This may include an institution code and an institution identifier.
Purpose	To allow a user to discover the location of the physical specimen to access additional information or to request access to the physical specimen.
Applicable standard(s)/recommendation(s)	It is strongly recommended to include a human readable term and a globally unique persistent and resolvable identifier from GrSciColl, GRID, ROR, etc.
Examples	dwc:ownerInstitutionCode : RBGE, dwc:institutionID : https://www.gbif.org/grscicoll/institution/0237598a-853a-492c-af74-a723fe251799; dwc:ownerInstitutionCode : MNHN, dwc:institutionID : https://www.gbif.org/grscicoll/institution/fe534fe7-dccd-4d79-8313-c11b2ea854ab; dwc:ownerInstitutionCode : BZ, dwc:institutionID : https://www.gbif.org/grscicoll/institution/6a6ac6c5-1b8a-48db-91a2-f8661274ff80
Required (Biological/Geological/Paelaeontological)	Yes (all)
Constraints	When there are multiple identifiers for the organization this term can be repeated as necessary.
Element specification status	agreed; accepted in specification
Notes

[RBGE-Herbarium]

From CETAF DWG meetings:

Should this Include the GRSciCol ID from GBIF or equivalent or should this be the dominant value? Stability issues over plain text name or ID.

[RBGE-Herbarium]

Chat from DiSSCo Prepare OpenDS session:

(Wouter)There is a DwC field for institution referent, which is dwc:institutionID. Issue with that field though is that it includes a recommendation to use an identifier from a collections registry, and cannot capture the type of identifier (GRID, ROR etc)
(Niels)TDWG suggests http://biocol.org/urn:lsid:biocol.org:col:34777. Why cannot https://ror.org/04py0zz23 be used?
(Wouter)@niels it could be used. It is just against the current recommendation.
(Mathias)Even if we don't have institutionIDType, the use of institutionID will resolve quite a bit of ambiguity already compared to institutionCode.

[akoivune]

How this institution code should be used if the physical location of the specimen is different from the institution that is maintaining and storing the data related to it?

[dagendresen]

dwc:institutionID (and dwc:institutionCode) identifies the institution holding the physical specimen, not the custodian of the data. Would not the custodian of the data record be sufficiently described by the dataset metadata?

@wouteraddink why do think that using e ROR identifier as dwc:institutionID is against the current recommendation? The term definition only says the recommended best practice is to use an identifier from a collections registry such as the Global Registry of Biodiversity Repositories

[wouteraddink]

@dagendresen, it could be used, but it is against the current recommendation as ROR is not a collections registry. In my opinion the recommendation should change. ROR is a resolvable persistent global identifier so should be perfectly fine, encouraged even and not be against the recommendation.

[dagendresen]

I see: the words collections registry. Thanks! Completely agree that a ROR ID is much more useful than GrSciColl ID/code.

[tucotuco]

Sounds like a quite reasonable request for change to me!

…

On Fri, May 7, 2021 at 1:40 PM Dag Endresen ***@***.***> wrote: I see: the words collections registry. Thanks! Completely agree that a ROR ID is much more useful than GrSciColl ID/code. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#11 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AADQ724WMXHCIVNOCRIRGXLTMQJY7ANCNFSM4WKK2UJA> .

[matdillen]

First attempt at a revised definition after discussion on June 3:

Definition: A unique identity for the specimen's curating institution. Whatever identifier the institution offers to uniquely identify itself, as long as this identifier allows a user to locate (contact?) the institution.

Recommended practice is to provide an ROR identifier. If this is not possible, it is recommended to provide an identifier from another registry, including the registry itself if it is not apparent.

[smrgeoinfo]

If this property is identifying an institution in a particular role (curation), I'd suggest naming the property 'curatingInstitution' so its clear.

[wouteraddink]

I think the role is better defined as data custodian then as curation, also given that we aim to enable community curation of the data. Also note that we now think of using Wikidata for situations where an institution cannot get a ROR. We may add WikiData as possible known referent.

[emhaston]

Discussion from TDWG Task Group meeting (5 May 2022)

Two parts - code + referent
Do we need the referent?
Referent beneficial for machine readability, disambiguation
However adds another layer of complexity in what should be minimum information about a digital specimen
Keep institutionID as it stands, consider institutionCode for perhaps MIDS level 2…?
Maybe separate into two separate elements; one for code, one for the referent?

Based on the discussion at this meeting, the Institution element has been separated into 2 elements. The InstitutionID element was agreed to be included in MIDS-1. More discussion is needed for InstitutionCode but there was a recommendation to include this at MIDS-1 too.

[emhaston]

Purpose: to provide a machine readable identifier for where the specimen is held.

[matdillen]

What is the importance of institutionCode if an institutionID is available? I think we just need the ID.

[emhaston]

We may want to simplify this term and provide more guidance in the recommendations and mapping.

Proposal: To change the name of this MIDS element to Institution with the following definition and recommendation:

Definition: A term to indicate in which institution the specimen is held. This may include an institution code and an institution identifier.
Recommendation: To include a human readable term and a globally unique persistent identifier from GrSciColl,
GRID, ROR, etc.

DwC mapping: dwc:ownerInstitutionCode; dwc:institutionID

[tucotuco]

Hi folks,

Just a point of information on the Darwin Core mapping. The term dwc:ownerInstitutionCode is meant specifically for ownership, NOT necessarily where the specimen is held. The term for where a specimen is held is dwc:institutionCode which is supposed to be the same institution as the one referenced in dwc:institutionID.

[emhaston]

Thanks for the correction. dwc:institutionCode it is.

[PietrH]

Definition: A term to indicate in which institution the specimen is held. This may include an institution code and an institution identifier. Recommendation: To include a human readable term and a globally unique persistent identifier from GrSciColl, GRID, ROR, etc.

The way it's worded now we recommend an institutionCode and an institutionIdentifier, Is this intentional, or is either of these sufficient? Or do we prefer the identifier over the code?

We should strive to provide a good example as well. I can certainly come up with some institutionCode/institutionIdentifier examples for our Meise Botanic Gardens herbarium dataset; BR or https://www.wikidata.org/wiki/Q55829049 or https://www.wikidata.org/wiki/Q3052500 or https://ror.org/01h1jbk91 or http://biocol.org/urn:lsid:biocol.org:col:15605 or MeiseBG, the list goes on!

[emhaston]

The recommendation of both is intentional with the idea that we would be recommending a human readable and a globally unique identifier. There has been strong recommendations for both of these pieces of information for different use cases and I think that we should be recommending both. However, both are not required in this proposal and either one would potentially satisfy the element of Institution.

[emhaston]

Task Group meeting (2 June 2022)
Notes for this element:
Institution - Simplify institution to a single element with a recommendation that people include a human readable term and a PID.
Risk that institutions will not use this properly if it is just in recommendation
Would need very clear recommendations and examples
Difficult to manage the requirements/quality within this framework
Change to Organisation (Schema.org)

Decision:
To change name to Organization (maps exactly to schema.org/Organization.
To include in all MIDS levels
To include as a 'generic' element in MIDS-0 and MIDS-1 with a recommendation to include a human readable code and a globally unique persistent identifier.
Further discussion required for MIDS-2 and MIDS-3

[emhaston]

Add wikidata mapping.