Home

Welcome to the PiCo wiki! Here we post the minutes of the last meeting, which was: April 2026.

During the last meeting, we focused on questions that were raised in the Github.

The most prominent discussion focused on what can and should be described using roles. During our disccusion, we stressed that:

• Roles are in general the worst place in the Persons in Context model to describe information. Information from roles that describe an association between two person observations should be modeled as a relation, as this allows for easier reuse. Such relations can either be a familial relationship defined in the Persons in Context model (https://personsincontext.org/model/#Persons-in-Context---Object-Properties) or a non-familial relation defined by a domain or data provider.
• Except for familial relationships (father, mother, child) or when definitions of roles are meaningless (registered, observed), data providers are free in their use of picom:roles and can either use strings or the predefined roles from the PiCo terminology.
• The current list of roles is a heritage of the A2A model, which allows an indefinite number of roles, leading to more than 20.000 roles in practice. For the Persons in Context model, the A2A roles were standardized and reduced to a brief list of person roles that are absolutely necessary to understand why a person observation appears on an archive component. As a result, the list of roles will not be extended unless historical sources require - that means, cannot be understood and queried without - a specific role.

Besides roles, we discussed maintenance​ on the Persons in Context model.

• We briefly discussed the modeling of places. No changes to the model were required, but we confirmed that our approach aligns with the schema.org application profile of the NDE. A few minor adjustments will be made to PiCo to ensure alignment with that application profile.
• Resource descriptions, data checks, and examples are now validated separately, which can lead to incongruencies. During the next update of the documentation, we will tackle this technical issue.
• Later this year, we will start working on a general SHACL for the Persons in Context model. The goal of this model is to perform a minimum data quality check on all historical sources. We will not provide source-specific SHACLs, but will showcase the SHACLs that have been developed by the PiCo community. The next meeting will be in April 2026.