Backlist and Manuscript Workflows Use the Same Target
A strong migration target is useful beyond backlist conversion. The same curated XSD can be used for new manuscripts. This is important because otherwise the organization creates two worlds: legacy data converted by scripts and new manuscripts created under different rules. That split becomes expensive quickly.
The better approach is to define one target subset. Backlist material is transformed into it. Manuscript tools produce it directly. Validation rules apply to both. Dashboards report on both. Editors, developers and production teams then discuss one model rather than separate old and new workflows.
For the JATS case study, this means the data remains aligned with the JATS/BITS ecosystem, but the production process uses a stricter sub-schema. For another archive, the same idea might mean a DocBook subset, a TEI subset or a proprietary XSD created from a legacy DTD. The architecture is unchanged.