State diagram for the new volume workflow: JSON-LD first or HTML first (basis for #19)

[WolfgangFahl]

🗣️ Wolfgang Fahl → 🤖 Agent/Guido: For #19 we need a state diagram, either as PlantUML or plain Graphviz. It should show the options: we can start with JSON-LD and create the HTML from it, with the JSON-LD embedded via semantify3 (JSON-LD is what we fence), or we can parse an HTML and create JSON-LD from it and then do the round trip. We need to discuss where the code goes: currently we have cvb.bitplan.com and ceur-spt, which are both quite heavyweight. Would it be possible to refactor properly later? But first let's do the state diagram as a basis for discussion.

[WolfgangFahl]

✍️🤖 Agent/Guido: state diagram, source docs/workflow/NewVolumeMetadataStates.puml

Sources: HOWTOSUBMIT and the Vol-XXX/index.html template for today's HTML-first entry; #16 for bootstrapping; the workflow and commit chain from the task force meeting of 2026-09-22 for the states from submission to downstream.

Agent reading, not from a source: the "Round trip checked" state with its extract-and-compare step; the "Rejected" state; placing "volume number, URN, publication date set" in the editor's check.

Open points: the question marks are taken over from the meeting notes (preview via git branch, dblp/GND/TIB passive). Transition labels are operations; which component implements them is open. semantify3 0.0.5 extracts yaml and sidif fences; a JSON-LD fence needs handling of its own.