google / schemarama

Schemarama is a project exploring standards-based validation for structured data, especially Schema.org.
Apache License 2.0
124 stars 22 forks source link

Figure out best pattern for linking from bits of ShEx (smaller than shapes) to documentation URLs #40

Open danbri opened 2 years ago

danbri commented 2 years ago

For a concrete example, consider https://github.com/google/schemarama/blob/main/demo/validation/shex/specific/ServiceB/Dataset.shex

which was in fact based on Google's rules for describing datasets to Google Dataset Search. This all got anonymized for the opensource launch here for simplicity.

Here's a fragment of the ShExC:

<#ValidServiceBDataset> @<#ValidSchemaDataset> AND EXTRA a {
    schema:description .
    // techdoc:url "https://schema.org/url"
    // techdoc:description "A short summary describing a dataset."
    // techdoc:identifier "error";

    schema:name .
    // techdoc:url "https://schema.org/name"
    // techdoc:description "A descriptive name of a dataset."
    // techdoc:identifier "error";

    schema:alternateName . +
    // techdoc:url "https://schema.org/alternateName"
    // techdoc:description "Alternative names that have been used to refer to this dataset, such as aliases or abbreviations."
    // techdoc:identifier "warning";

    schema:creator . +
    // techdoc:url "https://schema.org/creator"
    // techdoc:description "The creator or author of this dataset."
    // techdoc:identifier "warning";

    schema:citation .
    // techdoc:url "https://schema.org/citation"
    // techdoc:description "Identifies academic articles that are recommended by the data provider be cited in addition to the dataset itself."
    // techdoc:identifier "warning";
...

It might be that the linking from bits of ShEx to supporting documents is better supported by out-of-band metadata, rather than inline. To be discussed!

danbri commented 1 year ago

@ericprud has mentioned "shape paths" in this context.

out-of-band is important...