Indeed on the tricky bit. But even so, some dynamics might be discernible in the mess. Such as... consider the set of skeuomorphs defended by the claim to fun

Indeed on the tricky bit. But even so, some dynamics might be discernible in the mess. Such as... consider the set of skeuomorphs defended by the claim to function that they aid by minimizing some switching cost. That defense has something akin to a half-life, when there exists some set of transition steps that takes the design of the skeuomorph, imperceptibly at each step, toward some more efficient design. So indefinitely using the "minimizing switching costs" rationale as giving "function" wouldn't hold up in such cases. Then it seems reasonable to say that any costs which arise from not having moved such skeuomorphs along that path of imperceptible transitions in a timely fashion would be a indefensible waste if that was the skeuomorph's only claim to functionality.