--On Monday, 26 March, 2007 23:21 +0100 Tony Finch
<dot@xxxxxxxx> wrote:
> On Mon, 26 Mar 2007, John C Klensin wrote:
>>
>> In particular, none of [the "social network" tools] permit me
>> to maintain, with a single nominal identity, different
>> circles of acquaintances for different purposes and with
>> different trust and influence relationships between and
>> within each, an issue that was clearly understood in the
>> literature by the time I started reading it in the first half
>> of the 60s.
>
> LiveJournal allows you to subset your "friends" to provide
> relatively fine-grained access control to your posts, but it
> doesn't let you encode the more complicated relationship
> topologies that you're hinting at. In practice most people
> can't be bothered with complicated privacy arrangements and
> stick with friends-only or public in almost all cases.
I would have phrased that, at least in part, as "the human
factors/ human interface considerations associated with managing
a large number of relationship categories in an explicit way are
very hard". The difference is that we know, from years and
years of research on contact networks, that people really do
perceive these differences. A simple experiment in which a
questionnaire is administrated with collection of more or less
personal questions (the ones used by the popular social network
or dating sites would do reasonably well) are asked against
categories of "would tell... { parents, spouse, boyfriend,
girlfriend, best friend, lover, office mate, fellow employee,
boss, subordinate, person met in park, person met in
recreational activities, work-related acquaintance,
religion-related acquaintance, person met on net, ... }" will
turn their existence up in a hurry, even though it might not be
helpful in actually building category definitions and rules.
> Facebook has an intermediate level of locality-based
> "networks" which let you reveal more to people physically near
> you (at the same university) than to the unwashed masses, but
> still keep things back for your friends.
>
> http://nymag.com/news/features/27341/ is an interesting
> article about young people's attitude to privacy.
Nice piece. Thanks. FWIW, in the most recent round when I was
trying to do this sort of stuff seriously, the key issue wasn't
privacy in the usual sense but interruption and
attention-getting. So the key questions didn't pivot around
"who would you tell..." but, e.g., around "who are you willing
to have interrupt you at 3AM and who would you prefer to send to
voicemail". And the problem arises if neither answer yields a
null set and your caller definition is such that "calling
number" is not an adequate filter. If, e.g., the answer to the
"3AM" question is "no one", then there is an easy solution. But
the mixture usually requires that the phone ring and you or your
agent decide whether to pick up.
best,
john
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf