1. Paul Burchard's point about the social contract between
a web site and its visitor is very well made.
2. The BACK and FORWARD functions in the behemoths that we call
browsers are understood in this contract to play the role of
shuffling pages in the visitor's shopping list. No HTTP
transactions, including HEAD, should result from either BACK or
FORWARD except when reasonable memory usage in this regard (about
which the user may have configuration options) has been exceeded.
Even originally-provided meta-information is irrelevant for this.
(Yes, there is a scenario where a browser stays up while the user
goes on vacation ... Forget it! The user also has a RELOAD
funtion. Less is more!)
3. There is a big difference between a web page that provides HTTP
GET for shopping cart additions and deletions and one that
(against the contract) provides HTTP GET for checkout. (I am
tending to view this question on shopping cart additions and
deletions as a gray area. If additions and deletions were handled
client-side, it would not be gray at all.)
4. JavaScript has been so misused (and someone here would seem to be
proposing further misuse) that it causes me as web-visitor to
avoid entirely web-keepers who use it. I would be surprised if
anyone savvy enough to develop client and/or server tools is
foolish enough to visit strangers with it enabled. (Those who
would try to disable the BACK function can try to disable many
other functions.)
5. It was also the case in the early 90's that gopher requests did
not necessarily address static content. This idea is *old*.
-- Bill
