RE: full integrity?

From: Randall Gellens ^lt;rg+ietf@qualcomm.com>
Date: Thu May 23 2002 - 17:30:35 EDT

At 2:00 PM -0700 5/22/02, Randy Bush wrote:

> > The example I gave was in the context of some protocol where geopriv
>> is used, I ask you for something, before you reply you ask me for my
>> location. I don't want you to know where I am. I can either send
>> you a vague location, or I can send you multiple locations, one of
>> which is correct.
>
> or i can tell you to stuff it.

Well, of course, you can decline to disclose the information, but
that really is besides the question I tried to raise. In my example,
there is an assumption (and I apologize for not making this more
painfully clear) that I want the service, and in order to get it I
need to disclose my location. If I don't want the service (or if any
disclosure at all isn't worth it, or if any disclosure within the
required accuracy isn't worth it), then my question doesn't apply.

If I do want the service, and if I need to provide location within
some minimum accuracy in order to get it, one possible approach is to
say "OK, here are x locations as accurate as you wanted. Give me the
service for all x."

As others have pointed out, providing this capability naturally adds
some measure of complexity to our object(s). As a general principle,
simpler, less capable, and fewer options are better. As long as it
isn't too simple and too incapable to be useful. So the question
remains: is this capability required?
Received on Thu May 23 17:31:59 2002

This archive was generated by hypermail 2.1.8 : Thu Jan 22 2004 - 12:32:23 EST