Ian,
I don't quite see the difference between the WPAD case you cite and the
case where the
IS Department pre-configures my browser with a PAC URL. It seems to me that
in both cases the existence of the proxy is similarly unknown to the
content-consumer (and upper layers of the browser). In both cases it is
possible (I assume this is true for a WPAD-enabled browser?) to determine
the use of a proxy if one takes the trouble. In summary I'm not convinced
that auto-discovery introduces "transparency".
Dave
At 17:06 23/06/99 GMT, Ian Cooper wrote:
>On Tue, 22 Jun 1999 01:29:11 -0700, Josh Cohen (Exchange) wrote:
>
>>in the HTTP draft, the term "transparent", I beleive,
>>is meaning "semantically transparent".
>>In cache parlance, "transparent" seems to mean
>>"network transparent".. we should probably come up
>>with a good word for the wrec version of transparent
>
>Do we really mean (in our own language) that transparent means
>*network* transparent, or do we mean content consumer transparent?
>
>Consider a WPAD enabled environment: the content consumer or
>high-level parts of the client are unaware of the existence of the
>proxy (so long as the proxy is semantically transparent) and yet no
>clever trickery is going on.
>
>I agree that we use the terminology differently to HTTP[0], but in
>understanding our own definition of a "transparent proxy" we must
>understand what we mean by "transparent" in the first place.
>
>So, does anyone disagree that a proxy that is used by the
>configuration through WPAD is "transparent"?
>
>
>The current draft talks of "transparent proxy_ing_" - which might help
>differentiate between the HTTP definition. However, that difference
>is at the expense of playing word games in English, and I'm not keen
>to do that (which is why I want to try and get this sorted).
>
>It seems that, as a community, we are saying that our definition of a
>transparent proxy is one that is fed (by whatever method) by some form
>of traffic interception. If we take a more generalized approach, that
>means a transparent proxy is simply one whose presence is unknown to
>the content consumer. Yes? No?
>
>
>[0] In cache-speak a transparent proxy is going to have to do some
>special stuff to be semantically transparent. So perhaps it is
>appropriate to annotate the HTTP definition with a bit more
>clarification[1].
>
>[1] I _think_ I've managed to work out the fundamental difference. In
>the HTTP spec. "transparent proxy" is a verb-noun phrase. In our own
>language "transparent proxy" is a noun. If I'm right on that then
>does this help our understanding any further?
>
>
>Ian (again requesting comments on the draft in general)
>
>
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:26 MST