Jeni's Musings

Since XML and LMNL are completely disjoint both syntactically (LMNL uses [ and {, whereas XML uses & and < as markup signals) and semantically (LMNL has atoms, ranges, and annotations, whereas XML has entities, elements, and attributes), the best syntax for indicating dominance in LMNL documents is probably XML. Stripping LMNL markup from a LMNL document is extremely easy, allowing an XML parser to do its job. XConcur is about overlapping dominance, which is in my view quite disjoint from LMNL’s overlapping containment, so adding LMNL markup to an XConcur document would also be a useful thing in some circumstances.

Alternatively, one could devise a hybrid parser and data model, since both markup metalanguages do share the concept of characters. Such a parser would provide a LMNL range set whose elements represented both LMNL range markup and XML element markup, but also contained parent/child links for the XML elements. If it understood XConcur syntax, there would be multiple roots. XML attributes are a degenerate form of LMNL annotations, so there’s no problem there.

I don’t say this is ideal, but at least it’s straightforward.