On Fri, 3 Oct 2003, Abbie Barbir wrote:
Not necessarily. For example, if I know that my ISP is logging all
HTTP request bodies to be sent to FBI or a marketing agency, I
want to tell my ISP to bypass that feature. An OPES-compliant ISP
will honor my request if possible.
I recommend that you change ISP or better move to another country.
This is not what we are asked to do here by the IAB.
IMO, this is _exactly_ what IAB was (or should have been) concerned
about: corrupted or misbehaving OPES intermediaries. Clearly, it is a
matter of IAB RFC interpretation, but I would rather err on the safe
side than narrow down IAB implications to a trivial case.
Even today, content is being composed from dozens of sources and
pieces. An all-or-nothing approach would more and more often mean
"nothing". Not just will you strip the ad or weather info. You will
not get the news page at all. The URI-based approach lets you get the
page you want, with some missing information.
True, except we cannot tell whether it is the first or any other
OPES processor or callout service that will honor the bypass
instruction. We should not tell an OPES system exactly how to
implement bypass.
It is clear from (3.3) it is client centric (It is the 1st OPES
processor that will get the request)
While 3.3 indeed applies to content, it says nothing about the OPES
side (client or server) or the OPES processor number. While it is the
1st OPES processor that will get the request, it may not be the 1st
OPES processor that will know whether a non-OPES version is available.
IAB is concerned about some minimal functionality. We can get a
lot more, at about the same cost. We do follow what IAB 3.3
suggests, but we go further.
U can if u have the time. I suggest that u do the minimum for now
and get the OCP/HTTP drafts done, otherwise, just packup and go
home.
Agreed. That minimum, for me, is a URI-based interface with a wildcard
option and appropriate defaults.
3. Non-blocking MUST be Done in-band.
Do you mean "and only in-band"? What does this requirement buy
you?
What I mean, just ride on tracing (one extra header all what we
need)
Not sure I follow. Traces go in the opposite direction of bypass
requests so you cannot reuse them for bypass. I am not suggesting that
we specify an out-of-band mechanism, but I do not see any reason to
prohibit one either.
Alex.