> > The trouble is that <repository> is given without the '?'
> > quantifier thus making it mandatory if any other element from
> > sequence is present. This is confirmed by the validator: as long
> > as there's a <repository>, anything goes. I don't think this
> > omission of the '?' quantifier in ROMA is intentional, though.
> Ah, I missed that one. This indeed looks like an oversight in ROMA.
No, I don't think so. My recollection is that the manuscript experts
on that task force believed that a repository was required. Any of
the other MSsers care to comment?
(It would actually be an easier content model to maintain if
repository were also optional, because then what is now a reference
to a model class followed by four references to specific elements
could become one reference to a specialized model class.)