[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: The Carl & Bob show

On Thu, 20 Nov 1997 jar@ornl.gov wrote:

-> Bob said:
-> If there should just happen to be a Robert R. Jueneman who is a sheep herder
-> in New Zealand,  in all probability that individual is not likely to be
-> confused with the fellow who writes all of the quasi-legal stuff on several
-> PKI e-mail lists. Of course, as international electronic commerce becomes
-> increasingly common, that degree of global unambiguity will become
-> increasingly important.
-> ------
-> Gee, Bob I get about one E-mail a week for a Jim Rome who is a sportscaster for 
-> ESPN2 in San Diego and I am a scientist in Tennessee...
-> Jim Rome (the scientist)

I think that both are correct and this is a conclusion by itself.

Perhaps we can agree that any naming scheme (call it URL in http, e-mail
in SMTP/POP3, DN in X.509, key-fingerprint in PGP, key in SPKI/SDSI, etc)
has its scope and meaning entirely defined by the policy and interpretive
values of the issuer, but is perceived as projected by the observer in its
own reference frame.

For example, multiple SPKI public-keys may look different for an observer
but may be operationally equal to an issuer that calculates K mod n for a
pre-assigned n and uses only one private-key. 

Thus, "names" are not absolute references by themselves, in naming
schemes. One must add an "observer effect", for which one must also take
into account the interpretive value as seen by the verifier of the name,
which is an independent agent -- acting as an observer.

So, "Everything said is said by an observer" reflects the fact that the
observer (ie, the verifier) is the central entity here, so name
interpretation must reflect the occurence of any number of observers, all
in their own reference frames, all different from the issuer. And, all
intrinsically valid in their own domain -- even though possibly ambiguous
in other domains, or, the other way around ;-)
So, my opinion is that both Bob & Jim are correct. Perhaps ambiguity would
be better solved by taking it into account, instead of trying to iron it

Further, ambiguity should be actually seen as an artifact of our measuring
process because if we change the measuring process the ambiguity changes
and may even vanish. 


Dr.rer.nat. E. Gerck                        egerck@laser.cps.softex.br
P.O.Box 1201, CEP13001-970, Campinas-SP, Brazil