On 24/02/16 11:22, James Cummings wrote:
If I don't hear anything soon confirming/challenging https://github.com/TEIC/TEI/issues/564 then I'll be adding an @odd attribute with a data.pointer datatype.
Lou makes the (quite sensible) argument that instead of adding an @odd attribute we should create a <schemaRef/> element (with presumably a @target). This would enable a better content model of saying you could have either <schemaRef/> or <schemaSpec> (which is already allowed inside encodingDesc). Originally this issues was to add @methodology to <editorialDecl> but given what want to be tracked it was thought that this should be @odd instead, but this makes more sense to me (and StefanieG had also commented as much on the issue) to be further up on <encodingDesc>. Likewise, Lou's suggestion for a choice between either a <schemaRef/> or <schemaSpec> only really makes sense in <encodingDesc> where the latter is already allowed. (Just noting the slowing changing nature of this issue and why I've updated the title of it.) I like Lou's idea and am for it. (And could potentially start to envision where having a <schemaRef/> might eventually be useful in other places) However, adding an element seems a bigger step than adding a new attribute so wanted to return to council for any discussion before proceeding. -James -- Dr James Cummings, James.Cummings@it.ox.ac.uk Academic IT Services, University of Oxford