that person is a threat. Is it someone that it’s
all right to talk to? If the person, once recognized, is assessed to not be a
stranger, then the conversation can begin. Otherwise, an alternative course of
action is called for. We might tell the child to “Be aware!” or we might
suggest a stronger reaction; “Run away!”
Thus far, our
defined steps aren’t too difficult, except for the points that we haven’t thus
far considered. We haven’t yet defined what it means to recognize someone. Nor
have we said how we determine that they’re not a threat. Well, good grief! These
seemingly simple concepts called trust and policy, along with an equally simple
concept called a protocol for implementing trusted policy, suddenly offers the
prospect of getting a bit out of hand. Trust us, it gets worse! For example, we
certainly need to more fully consider what trust really means. When we ask for
trust, as we’ve just done, how does that relate to subsequent interactions that
are driven by policy? For an answer, it will help if we first delve a bit
deeper into policy itself.
Let’s consider
in some detail the implementation of policy that lets us visit a friend in her
apartment. It should now go without saying, but the point is so central to this
particular protocol that we have to say it again, “Don’t talk to strangers!”
Thus, our friend in the apartment must, or at least really should, establish
some degree of confidence that entry of the visitor is desired, or at least
benign, before opening the door. Few will open the door for a person they
perceive to be a threat; that’s the whole point of the door, the lock and the
intercom system. Through these mechanisms, the friend is offered the
opportunity to divine the identity of a visitor, at least to such a level as to
ascribe some degree of assurance that the visitor is not a threat. Now, the
simple policy conveyed by a sign that was marked PULL has rather quickly
encompassed many of the characteristics of a social ecosystem, a
concept that we’ll get into in some detail within this book. The
characteristics that we’re particularly interested in comprise the
infrastructure for general interactions, and how they enable a very specific
form of interaction which we’ll term a transaction.
This is an interaction for which we can uniquely define and apply policy, from
a well defined beginning to a well defined end.
The moderately complex
protocol we’ve just considered is rather ubiquitous in most urban environments.
Consequently, the policy and its subsumed protocol are disseminated through
social convention. We might also consider another example, the policy for
opening the door of a tenant safe in a hotel room. This procedure is probably
unusual enough for the common traveler, who might well see lots of different
variants of room safes in hotels around the world that it’s useful to have the
protocol written down on an instruction sheet. Generally the instructions are
presented in several languages in order to accommodate the widest variety of
travelers. For most such safes, the protocol begins with the instructions to
reset the safe to a known starting state and to allow the guest to enter a code
that only she knows. Once the door is closed and locked, the safe will open
only when that code is entered.
So, thus far
through our examples we’ve discerned two distinct mechanisms for establishing
trust within an interaction. In the first, the friend hears the voice of a
visitor at the door and, based on purely subjective means of identification,
decides whether a state of sufficient trust has been established so as to
warrant opening the door. In the second, the safe door sees the appropriate
code entered such that it can also determine that a state of sufficient trust
has been established, based on the purely objective decision rules built into
its control circuitry, so as to warrant opening the door. In either case, the
protocol was an element of a system through which trust could be established in
order to support the occurrence of a transaction. Trust is our measure of
confidence in the validity, consistency and probable consequences of the
specification and application of policy. To elaborate slightly, we can say that
within an environment in which an
|