I think Brian hit it right on -- request-id, time, referer, and user-agent.
Brian gave reasons for the first two. Referer seems to be the equivalent
of Nielsen demographics in some providers' minds -- especially those who
are paying for referring links. User-agent not only gives 'browser
capability' information, but also often platform information.
If the proxy services more than one domain (that's possible, yes? -- maybe
one of the proposed SOAPs from a few months ago), the last half of
'Forwarded' would also be frequently-requested.
Another way of looking at the list above is that it describes what
commercial providers currently get from most usage-analysis reports. If
you reduce the information in batch reports from proxies beyond this, I
suspect what you're calling 'open rebellion' will occur. The question will
be, why are we getting less information now than we were before? Having
proxies lock them out for non-compliance is a great idea, but I suspect AOL
or CIS would have to do so before most businesses would pay attention.
>> *every* client of ours wants stats as to the busiest time
>> of day for their sites
>
>I don't get it....isn't the point of electronic commerce to break
>out of the constraints of space and time that limit ordinary
>commerce?
Back up a step. We tell them, "Now customers can buy your products
twenty-four hours a day." And they say, "Okay, how many products have we
sold between midnight and eight A.M?" To them, it is a matter of
justifying their web presence and the expense of providing it. This is
actually the same reason more than just hit counts are important to
commercial providers: electronic commerce prevents them from _seeing_ their
customers, so they want to know whatever else they can. I'm sure they
would all want From as well, but I'm more than happy to leave that out (he
says, deleting three pieces of junkymail from Usenet posts, and two pieces
of 'webmaster@' junkymail ..... not bad for a week).
Marc Hedlund <[email protected]>