There shouldn't be a verdict at all as well as there shouldn't be something like a correct or an incorrect answer. This is good for early learners burning the mid night oil, cramming for the exam -:)
I think the following line ( from that Siunia's link - p.30 ) describes it best :
"Provisionall conclusion
The fuzziness attached to the genitive of measure construction is not merely syntactic. It is clearly constructional."
Like I said earlier in this thread it's treated as an exception to the general rule.
Another thought on page 5 is very interesting:
"Jackendof’s valuable postulate“(…) fuzziness must not be treated as a defect in language; nor is a theory of language defective that countenances it. Rather, fuziness is an inescapable characteristic of the concepts that language expresses. To attempt to defiine it out of semantics is only evasion (Jackendoff 1983: 125)”.
I really suggest reading these 30 pages to everyone who’s not a grade-hound, every page but slowly and attentively and then draw their own conclusions. It shouldn’t be a brainstorming session. On second thought, for some it certainly will.