Open sjhuskey opened 6 years ago
I feel as if I am missing something here, but it may be worth remarking that, if enforced by the schema, this would make manual encoding much more difficult. Right now, I can use a fill-in snippet in TextExpander (or indeed Oxygen) to produce valid XML and only have to type the variants, but having to come up with an ID for every use of <app>
would prevent this.
It's definitely a trade-off. The concern is whether or not we want the individual app
elements to be citable in a CTS or CTS-like scheme. In that case, it's essential for them to have ids.
I've thought for a while that we might need both a work-in-progress schema and a final product schema. IDs are trivial to add with a script/XSLT/XQuery pass, but possibly too much burden on human editors. At the same time, there are scenarios where IDs are crucial, because the logic of the document demands linking between elements.
It might be helpful for those wishing to work with the data on their own to require
@xml:id
on every<app>
.