Yves, regarding your comment 227, I think you found a gap in the glossary module spec and that the spec does not currently capture the intent.
I do not believe that the intent is to allow for listing glossary items without associating them with local occurrences. The discussions always reflected strong consensus that the XLIFF glossary is strictly local and to be associated with local occurrences.
The reason that both refs are OPTIONAL in the module's elements' definitions is IMHO that XML Schema cannot express that *at least one of them* is REQUIRED
So summarizing the above, I believe that it is correct that both candidates and glossary modules are using the same referencing mechanism and that the normative intent should be disambiguated by adding a PR saying that at least one occurrence of the ref attribute is REQUIRED per a <glossentry>, no matter if on the glossentry itself or a child <translation> element.
Please note that we do not REQUIRE exactly one occurrence of the ref, this is to allow for adding translations and corresponding references once the target is populated.
All, Yves, please let me know if you agree with my interpretation of the intent and with the proposed way how to disambiguate it in the current spec.
If I do not hear alternative proposals during today and tomorrow GMT, I will turn this into a CFD.
Thanks and regards
dF
=======================
LRC | CNGL | LT-Web | CSIS
University of Limerick, Ireland
telephone: +353-6120-2781
cellphone: +353-86-0222-158
facsimile: +353-6120-2734