Conventions for sheet formats and semantics

tabby records can conform to a particular (set of) conventions that cover individual composition of a record, the semantic definition of its components, and/or the employed vocabularies.

Employed conventions are declared on a per-sheet basis, by including their label directly in the sheet name (see Sheet convention declaration).

Technically, a convention can be thought of as a (default) provider for any sheet component (context, overrides, JSON data, or even a sheet itself). This means that if a record does not contain any such sheet component the convention specification is checked for a matching component, which is then used, if it exists.

Conventions enable rather minimal tabby records. They can adopt semantics and override logic from a convention, and thereby avoid duplication across individual records. Only on load a complete record is built and reported. Individual sheets in a record need not use a single convention, but may mix-and-match from any available convention.

Built-in conventions

• Collection-of-files dataset (v1, tby-ds1)
• Scientific dataset (v1, tby-sd1)

Adding new/Updating built-in conventions

Labels of all built-in conventions:

• start with the tby- prefix. This is an arbitrary prefix that aims to minimize name conflicts with user-provided conventions.

• include a short name for the convention.

• include a version number. This version number is incremented whenever backward-incompatible changes to a convention specification are made.

Convention specifications are put into a datalad_tabby/io/conventions/<label>/ directory in the package source tree. Because the version is included in the directory name, it is ensures that there are no naming conflicts when versions are incremented, and previous specifications can remain available indefinitely.