Hi all, The comment 212 stated that ctr:revisions@ref was NMTOKEN in the space and xs:anyUri in the schema.
https://lists.oasis-open.org/archives/xliff-comment/201402/msg00012.html That has been corrected: they are both NMTOKEN now. However, the following text was added in the issue tracker: [same issue applies also to gls:glossEntry@ref, gls:translation@ref, and mtc:match@ref] So Tom changed the type for those two refs as well. But they were (and still are) IRI in the specification. Personally I think having NMTOKEN for match@ref would OK. But for glossEntry@ref, gls:translation@ref it would mean we can't point to content in the target. In any case: we have a discrepancy now. I think glossEntry@ref, gls:translation@ref and match@ref need to be set back to xs:anyUri in the schemas. -yves