Background: Les is a seasoned librarian and metadata specialist who works with Smithsonian Institution and is experienced in using Wikidata.
Goals: Use Wikidata and a local Wikibase installation to investigate using it as a constituency database across Smithsonian units.
Skills: Proficient in metadata modeling, Wikidata, and SPARQL, but has never worked with a new Wikibase installation.
Challenges: Wikibase is a new area, and installing a Wikibase to test is difficult within U.S. federal computing guidelines, in terms of operational security and permitted software.
I want toinvestigate using Wikibase as a possible constituent database system within Smithsonian
So thatinternally at the Smithsonian we might have ways to connect and disambiguate constituents (people, organizations, companies, etc) across units, with the possibility of providing this as a public interface for linked open data and federated search in the future.
Les familiarizes themselves with Wikibase by engaging with the Wikibase Stakeholders Group, and seeing what other deployments have been done.
2
Les works with Smithsonian colleagues in various libraries, with the Wikimedian at Large, the Digital Transformation office, and the office of the CIO on possible directions on how to proceed, especially since Wikibase is not on the roster of approved software systems allowed on federal computing infrastructure.
3
It is decided to try a pilot program with an outside contractor with Wikibase specialization, and to use an external service on cloud computing infrastructure.
4
After lengthy budgeting and procurement procedures, a vendor is selected and Wikibase is installed on a private cloud computing system that can only be accessed on the internal Smithsonian intranet.
5
Some delays are met when installing tools familiar from Wikidata use, such as Quickstatements or TABernacle. Many tools do not work out of the box, requiring custom bot scripts or using the API to work with data. Wikidata Query Service works, but only when completely logged into the Smithsonian VPN, and does not work via web proxy.
6
WikiNames is launched internally for testing with some significant questions related to modeling Smithsonian constituency data, and whether to follow Wikidata's particular choices.
7
Les convenes a regular meeting of Smithsonian participants to help populate and experiment with the Wikibase/WikiNames.
Example of WikiNames Wikibase instance at the Smithsonian Institution.
Work with Smithsonian libraries colleagues, Wikimedian at Large, the Digital Transformation office, and the office of the CIO on possible directions on how to proceed.
Wikibase is not on the roster of approved software systems allowed on federal computing infrastructure.
Wikibase Cloud, Amazon Web Services, GCP, others
External contractor
Pilot program engaging an outside contractor with Wikibase specialization.
Requires an external service on cloud computing infrastructure.
Installation
Wikibase installed on a private cloud computing system that can only be accessed on the internal Smithsonian intranet.
Lengthy budgeting and procurement procedures results in a vendor selected.
Enable the suite of familiar Wikidata tools to work, such as query interface, data ingestion, and data editing tools.
Many tools do not work out of the box, requiring bot scripts or using the API to work with data.
Some delays are met when installing tools familiar from Wikidata use, such as Quickstatements for data ingestion, or TABernacle for interactive data editing.
Create personas and possible scenarios for internal Smithsonian usage.
Create internal pages:
"Smithsonian WikiNames is a shared platform for the Smithsonian staff and associates for identity management and name authority for names—of people, corporate bodies, organizations, expeditions, people groups, personalities, and more."
WikiNames:Vision and Goals
Ongoing experimentation
WikiNames launches to internal Smithsonian users, with a proposed data model and encouragement ot upload data sets for real and fictional entities.
Open questions about modeling exist, and how Smithonian's proposed model fits with Wikidata. Questions about how to host images that are not from Wikimedia Commons, possibly using IIIF manifests from Smithsonian IDS.
Wikiconference North America presentation - https://wikiconference.org/wiki/Submissions:2024/LOFESQ_(Lots_of_Farmers_Empty_Silos_Quicker):_building_community_through_a_named_entity_Wikibase
