mjhawkins
commented
1 year ago I think that the best solution is to not have warnings for this in GS. That schema is merely about confirming that the XML has our shared understanding of how things should be coded. Warnings about X or Y not being processed in the current web application strikes me as a different order of validation.
Perhaps we need several orders -- ie. validation to ensure that TEI files conform to our shared understanding; validation that the file will work in the Cambridge Digital Collection Platform (or Platform X - implemented locally). Both of those would be mission critical. We'd only want the tests that ensure the item will work in the system.
Then, there's more strenuous validation that's intended to pick up code issues that don't cause problems or seriously affect the delivery of the item (we're not currently dealing with rend on element x; we don't process element y, etc, we aren't acting on the status attribute in revisionDesc's change element). These are issues that might not warrant any subsequent action (ie. the items work), but they might indicate areas of future development for a local implementation (maybe certain change status values should be displayed at the bottom of the document with a 'last modified' sort of message) or maybe we need to constrain status values because it suits our work editorially to know which items are in the first, second or third proofreading stage.
The consolidated schema currently warns about the use of rend on graphic:
The web site currently only supports rend attributes for hi elements. Using it on graphic elements is valid but will be ignored.
We use hi on surface/graphic to indicate orientation. Could this be made contextual (e.g. not applied to graphic elements which are a direct child of surface elements) or perhaps a candidate to be a rule specific to the manuscript catalogues?