Joe Touch wrote:
> But that would be outside of WREC; for WREC to suggest
> it as proposed, we'd have to do a little more work
> on it directly, no?
Common sense says that the WG needs to at least read and approve the
content before putting it's name on it. Not sure we are ready for that
as the problem of client access has to my knowledge not really been
discussed yet.
The basic issue dealt with in WPAD definitely belongs to the WG charter,
and as a protocol it belongs there no less than inter-cache
communication protocols or WCCP (which in part tries to address the same
basic issue as WPAD). It is a solution to a known problem most larger
organizations have to deal with; "how to get the clients to use the
correct proxy/cache for their network location". If this problem isn't
mentioned in the known problem draft then it should be added. Only
question is if WPAD is the best solution to the problem or if better
methods can be found. Personally I prefer it to redirection protocols
like WCCP by the simple fact that is does not change end-to-end
semantics of TCP/IP and yet provides a fairly detailed administrative
control, but both methods have their advantages and uses. The fact that
there already exists two vastly different approaches to the same basic
problem tells me that there quite likely can be found a more suitable
protocol/method which suits both audiences.
My vote is to have WPAD released as a informal RFC for reference until
the WG gets to the point of standardizing client access and
distribution, this is unless I have misunderstood something important
about the relations between WG character, standard track and informal
RFCs.. WPAD is after all implemented in at least one of the major
content clients (IE5), and can be deployed in most TCP/IP connected
networks as-is. I am a bit more restrictive on the WCCP draft as it is
more closely tied to a vendor, and operates at a very low level in the
network.
-- Henrik Nordstrom
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:26 MST