After extensive discussion of the date parsing issue, in light of the
agreement to include an “other” field containing unparsed date
content, Rintze and I settled on the idea of capturing the unparsed
content in two sub-variables to the date. These will be passed
through as prefix and suffix to the parsed and formatted date content.
Because the position of the unparsed parts is the same in all date
forms, and because formatting cannot be safely applied to them
(because the semantics of their content is unknown) we concluded that
they should not be included in the schema.
We propose the attached patch, which would implement this change by
dropping the “other” element from the schema. The new date-prefix and
date-suffix date parts would not appear in the schema, but would be
documented in the spec.