Structure

Documenting Collection Versions: Encoding, Formats and Identifiers

A collection version record can give several decisions a common home: the character repertoire, the chosen formats, the identifier scheme and the metadata profile. The purpose is a precise description of what a particular collection release means. A useful record would name each dependency, its version or dated documentation, and the part of the collection to which it applies.

A collection record should have room for those differences, rather than reducing every change to a single version number.

Character repertoires

The Unicode 18.0 announcement puts the encoded repertoire at 172,808 characters. Published on 16 September 2026, it describes a release containing character additions and code charts, revised data files and updated specifications governing fundamental aspects of text processing.

For a collection version note, those are distinct subjects: the repertoire used to describe the text, and the associated data and specifications. The historical Seal and Jurchen additions make the repertoire question particularly concrete for cultural material. The announcement identifies Seal as the largest addition and describes Jurchen as a cultural heritage script from China.

An encoding entry could therefore name the Unicode version and the historical scripts relevant to the material, with the associated data-file version recorded alongside it.

Format and accessibility profiles

The Library of Congress's Recommended Formats Statement for 2025-2026 assesses the potential of formats to support digital accessibility. The accessibility questions include text structure or tagging, identified page regions and headings, screen-reader navigation of tables, and forms that can validate entries.

The Library expressly does not require these features to be enabled for inclusion in its collections. Its evaluation recognises the capacity to support them as user expectations and communities change. The distinction is between a format's potential and a feature being present in the object.

A collection version record could give each its own description: the format chosen, its documented capabilities and the accessibility features actually enabled in the published material. The dated format assessment would belong alongside those entries.

Identifier migrations

Overture's identifier migration has a named release boundary: 2025-06-25.0. Earlier identifiers carried location and theme information; their replacements are random UUIDs. The documentation states that identifiers captured before that release will not resolve against current data.

A collection drawing on such data would need an identifier entry specific enough to distinguish the two schemes.

Scope matters within a release too. Overture's base-theme features and building_part features have deterministic UUIDs outside GERS, with no stability commitment. A version record should therefore identify the feature types covered by an identifier policy, as well as the policy's release boundary.

The same documentation says identifier-prefix filtering was never a reliable spatial query and directs spatial selection to the bbox column. A proposed record could distinguish the field used for spatial queries from the identifier used to refer to the feature.

Metadata application profiles

The BIBFRAME Interoperability Group's application profiles document common data practices by entity type. Its page, dated 2 May 2025, lists monograph and serial profiles expressed through Dublin Core Tabular Application Profiles, with conversion to Shapes Constraint Language for validation.

That provides a model for a more specific metadata entry than the vocabulary name alone: the application profile, the bibliographic format it covers and the validation representation. BIG's purpose is to make BIBFRAME shareable according to common practices. Its disclaimer also establishes an authorship boundary: the outputs are not official Library of Congress specifications.

BIG also states that it plans shapes for further bibliographic formats. A collection's metadata entry could name the profile actually adopted for that release, with prospective coverage recorded separately.

Across all four areas, the useful unit is a named decision with a defined scope.