I think the use of URIs for the values of the "Source" element would go
along way to addressing these problems - though of course yes, it would
be nice to have services that provide some metadata about what those
URIs identify (and/or deliver other related functions).
It seems crazy to me that (if I read it correctly) 9.2.1 Source has a
LangString datatype, with no guarantee of global uniqueness (MESH =
"Medical Subject Headings" or "My Esoteric Subject Headings" etc etc
etc), when, as Charles emphasises, disambiguation of the the value - in
a global context - is absolutely critical for semantic interoperability.
Whether that URI also de-references to a representation or description
of the vocabulary is a separate question; (it seems to me that) what
matters for Charles' purpose is that it provides a globally unique
identifier for the taxonomy (and for specific versions of taxonomies, if
those distinctions are significant).
Pete
|