Nor me. My understanding is that Concepts have the explicit name structure that they have due to more than just localization, and that it is an important feature of the system to be able to record exactly what concept name was asked and answered (eg. the specific wording) on a clinical observation. This is why, for example, Obs has a reference to valueCodedName in addition to valueCoded. As @dkayiwa pointed out, there may be value in trying to limit Concepts to clinical terms and value sets rather than a generic multi-lingual dictionary. Though I defer to @burke or @akanter on this.
Taking this a step further, one could ask why we have different types of metadata at all, if everything is ultimately able to link back to Concept. What design decisions led us to create distict “drug” table and independent Drug object rather to have Drug extends Concept and an explicit relationship between the “concept” and “concept_drug” table? The answers to that might be informative.
Mike