Expanding the Manifest Schema

[AdamSobieski]

Hello. I would like to share an idea that the manifest JSON resources in PWEB archives could be expanded to enable scenarios beyond those where index.html resources were the main resources in the archives.

Presently, PWEB archives' manifest resources have required fields:

spec_version
id
version
title
entry

and recommended fields:

description
author
created
icon
permissions
rights

In the specification, the entry field is described as typically being index.html.

One possibility is that a new field, entry_type, could indicate the MIME type of the entry resource.

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry_type": "text/html",
  "entry": "index.html"
}

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry_type": "application/xhtml+xml",
  "entry": "index.html"
}

and, for a non-HTML entry resource:

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry_type": "text/turtle",
  "entry": "index.ttl"
}

Another possibility is that one could expand the JSON schema to enable providing types for entry resources and to enable providing multiple alternative entry resources, enabling a form of content negotiation from inside of PWEB archives.

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry": [
    { "url": "index-en.html", "language": "en", "type": "text/html" },
    { "url": "index-fr.html", "language": "fr", "type": "text/html" }
  ]
}

and, for non-HTML entry resources:

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry": [
    { "url": "index.ttl", "type": "text/turtle" },
    { "url": "index.rdf", "type": "application/rdf+xml" }
  ]
}

What do you think about these ideas? Thank you.

[omscse]

Thanks @AdamSobieski ,
How will the non-html resource be used in PWEB?
How do you think PWEB viewers will handle the non-html resources?

PWEB Viewers are expected to render web content (html/css/js etc). I am curious to know the non html behavior that you are looking for.

[AdamSobieski]

@omscse, for starters, I'm thinking about argumentation-related scenarios. In addition to HTML-document resources, argument interchange format (AIF) resources could be entries. These entries could potentially reference other resources, e.g., illustrative multimedia resources and evidential resources (e.g., documents, spreadsheets, tables, charts, event logs), bundled alongside them in PWEB archives.

AIF arguments can be represented using RDF-based formats. So, initially, I'm thinking about Turtle and JSON-LD for entries, in addition to HTML.

Another potential use case would involve including data-table schemas for one or more data tables (e.g., text/csv, text/tab-separated-values). As envisioned, a metadata schema resource, CSVW metadata (application/csvm+json), describing the tables' schemas would be the entry resource, accompanying the one or more data tables bundled together with it in a PWEB archive. Perhaps, one day, Web developers will be able to download PWEB archives to efficiently create and/or populate local database instances using the IndexedDB API? See also: SQLite3 database files (application/x-sqlite3).

Another potential use case would involve including rules, grammars, schemas, or ontologies with which to verify one or more data resources. Data resources could be entries and, elsewhere in PWEB archives, there could be bundled other resources with which to verify them. For XML-based data resources, there is an xsi:schemaLocation attribute for referring to XML schema resources which could be stored in PWEB archives. For RDF-based data resources, there are dct:conformsTo and dash:shape properties for referring to other resources, e.g., SHACL resources. For many other data formats, PWEB archive metadata could be of use for interrelating data-resource entries to those accompanying rules, grammars, schemas, or ontologies with which to verify them.

Another potential use case would involve 3D models (e.g., model/gltf+json, model/xsd+xml, model/vnd.usda) which could, as entries, be accompanied by additional graphics, audio, animation, metadata, semantics, and/or accessibility-related resources. See also: Universal Scene Description packages (model/vnd.usdz+zip).

Another potential use case would involve using source-code resources or software projects as entries (e.g., text/x-python) such that one or more discussion transcripts (e.g., application/vnd.vcon+json) about the source-code resources could accompany them in a PWEB archive. As envisioned, dialogue moves would be able to include references to portions or selections of source-code resources.

Another potential use case would involve using markdown resources (text/markdown) as entries.

Interestingly, software systems (e.g., AI agents) could produce, store, load, transmit, and consume multiple resources bundled together into PWEB archives. Perhaps entry formats other than HTML would be useful for these scenarios?

Also, what do you think about the internationalization and localization and content-negotiation capabilities that would become possible by expanding the schema for the entry field in PWEB manifest resources?

{
  "spec_version": "0.1",
  "id": "org.example.minimal",
  "version": "1.0.0",
  "title": "Minimal Example",
  "entry": [
    { "url": "index-en.html", "language": "en", "type": "text/html" },
    { "url": "index-fr.html", "language": "fr", "type": "text/html" }
  ]
}

[BigBlueHat]

Y'all may want to look into the already standardized Web Publication Manifest format. It's lineage is EPUB related, but it could certainly be used for what you describe--and it already accommodates what the current schema here attempts.

Hope it's useful!
🎩

[omscse]

Thanks @BigBlueHat . PWEB addresses a different things.