Rationale behind storeKey?

Question: why is the request method used as input to the storeKey?

In HTTP, you can't cache different things based upon the method; e.g.,
if you cache a POST response (which is legal, just not implemented by
anybody AFAIK), that is cached for future GETs, not POSTs. HEAD could
theoretically be cached as something separate, but in practice it
seems to me that assuring that keeping acccounting for separate GET
and HEAD cache entries correctly (e.g., updating headers, etc.) is
more expensive than just computing a HEAD response based upon a cached
GET.

What am I missing?

Cheers,

--
Mark Nottingham       mnot_at_yahoo-inc.com