The key to living in a city and peacefully co-existing as a social
animal in tight quarters is to set a delicate balance of seeing and
not seeing. You take care not to step on the heels of the woman in
front of you on the way out of the subway, and you might take
passing note of her most excellent handbag. But you don't make eye
contact and exchange a nod. Or even if you do, you make sure that
it's as fleeting as it can be.
Where I live, there seems to be a division between "inside" and
"outside" spaces. In inside spaces, one does more than make eye
contact and nod -- we verbally greet each other (the old, the young,
and the especially polite will tack on a gender-based title to the
time-of-day greeting). In outside spaces, as a nod to pragmatism,
it's OK not to greet strangers (except maybe on Sunday mornings when
activity is sparse enough that the outside exemption no longer
applies, and greeting is again obligatory), but one should still
greet people one knows, by name.
Subways in general are crowded enough that they count as "outside"
space, but, just as with elevators, when people take a seat opposite
or next to you, greetings are expected, as are farewells upon
vacating the seat. (if anyone knows of special terminology for these
"inside" and "outside" spaces, please tell me)
-Dave
(also, I think there may be a little more going on with eye contact,
even in non-greeting societies. In order to move through a crowd, I
think it helps to very briefly make eye contact with people who may
be on a collision course, quickly breaking it by refocusing on one's
intended path -- when both parties do this, it's subconsciously
obvious what slight deviations need to be made)
:: :: ::
Social interactions don't like a lot of latency either
in audio or video (you notice the latency in calls
routed though geostationary comsats)...
Audio latency bounds are much tighter than video, and haptic bounds
tighter yet.
If our Ops/J ratio is horrible, it's because we're not doing
things right. We're a long way from KT ln 2, and even that
is not really a hard limit.
If I remember correctly, a little annotation:
"Ops/J" -- how much computing can we do for a given amount of energy?
"KT ln 2" -- a heat/entropy measure, showing the usual cost of
erasing bits
(not a hard limit, because if one can avoid erasing bits (via an
arbitrarily slow reversible computation?) theoretically one can
compute with arbitrarily small energy gradients)
