ACM is looking for an Editor-in-Chief (EIC)
for its International Conference Proceedings
Series (ICPS). ACM has initiated ICPS to
publish proceedings of high quality
conferences, symposia, and workshops
that are not sponsored by ACM or its SIGs.
ICPS has been in existence since 2002,
providing conference organizers a means of
electronically publishing proceedings that
ensures high visibility and wide distribution
(ICPS proceedings are made available in the
ACM Digital Library).
As a result of the sharp growth of the ICPS,
the ACM Publications Board has found
it necessary to initiate a more formal
editorial management structure for the
series to continue to ensure ACM-level
quality. The series will be managed similar
to other ACM publications and will have
an EIC who, along with the Editorial Board,
will control the content of the series. The
EIC will be further assisted by an Advisory
Board, which will include representation
from ACM SIGs.
ACM Publications Board has set up a
nominating committee to assist the Board
in selecting the ICPS EIC. The nominating
committee members (in alphabetical
Beng Chin Ooi,
National University of Singapore
The Open University, UK
Tamer Ozsu (Committee Chair),
University of Waterloo
Dan R. Olsen,
Brigham Young University
The nominating committee would like to
receive nominations for the EIC position
(self nominations are welcome). Please
send nominations, accompanied with
a short statement (up to one page)
justifying why the nominee is suitable
for this position, to Tamer Ozsu (tozsu@
cs.uwaterloo.ca) by October 15, 2009.
experiments were. Essentially, what
e’d done was create something that a
computer scientist would recognize as
a decentralized algorithm.
[CONTINUEd FrOM P. 112]
You’re referring to an experiment in
which Milgram asked a group of peo-
ple in the Midwest to forward a letter
to a friend of his near Boston.
Right. He gave them the man’s
name and mailing address and some
basic personal details, but the rules
were they had to mail it to someone
they knew on a first-name basis. And
no one had a bird’s-eye view of the network, but collectively people were able
to pass these messages to the target
How does your own work fit into that
What I did was use the ‘algorithm’
almost as a scientific probe of the structure of the network. I proposed a model
for social networks that allows people
to succeed at this kind of search, to land
as close to the target as possible. There
was a mathematical definition and linking trees and theorems. But the key
feature is that we should create links at
many different scales of resolution and
in equal proportions.
There are levels of resolution we can
use to think about the world. Geographically, there are people who are close
neighbors, people who are nearby, people who are in the same region and country, and so on. It’s important to have
links across all these scales. If we only
link locally, we can’t get messages far
away very quickly, but if we only create
links at long ranges, then, for example,
it would be very easy to get a message to
the Boston area, but there would be no
local structure to go the final few steps.
Have you been able to apply those find-
ings to more technical topics?
One area where they prove to be
useful is in the design of peer-to-peer
networks. If you think about what
a network is trying to do, it’s trying
to get information between pairs of
computers that need to communicate, and it’s trying to do this without any global knowledge of the network.
“A lot of the way
now is in a
it’s coming to us
we read, things
we see on twitter,
things our friends
You’ve also been working on a differ-
ent kind of Web search.
In the classical model of Web search,
some centralized thing like a search engine gathers up lots of Web pages, and
users come and issue queries and get
answers. A lot of the way we experience
information online now is in a different
form. It’s coming to us continuously, in
bite-size pieces, through our social networks—blogs we read, things we see on
Twitter, things our friends email us.
How should our search tools and algo-
rithms evolve to fit this new model?
One thing we should look at is a sort
of temporal pattern, a pattern across
time rather than across the network. So
measures like chatter, bursts of activity, and upward and downward trends
are going to be important, because they
help us organize this kind of information. You see this increasingly on sites
like Google News and Twitter search,
which are essentially trying to give a
time signature for certain stories.
Leah hoffmann is Brooklyn-based technology writer.