> What I hear is that we somewhere must have equivalence tables, regardless > of what charset we are using. how complex is the implied processing for a single lookup and what various places (may|must) it be performed? if it is complex, then either or both lightweight clients/resolvers or loads on servers become an issue. randy