Thanks — glad the word seek felt study-ready.
Yes on the important part: Forties and Muwatta already share one identity layer. Every Arabic word is a stable HUSX token id (uh:token:…). Timing maps and EN/UR glosses are independent sidecars that both key off those same ids; the core corpus text is never rewritten when we add a layer.
In practice today:
Glosses — one shared schema (husxGloss: 0.1) across Nawawi / Qudsi / Shah Waliullah and all 1,829 Muwatta reports: { tokenId → { en, ur } }.
Word timing — same token ids, different sidecar shape (start / end on each token). That layer exists for the Forties (where there is audio). Muwatta is text+gloss only so far, but when audio lands it plugs into the same token ids, not a second vocabulary.
So the Universal Standard idea here is: one token stream, many optional layers (timing, gloss, translation, refs…) — not one mega-file per collection.
