Granting a Second Life
Almost 90% of what we learn comes from reading, estimates say, and the path that
online information takes to your brain is less direct than most. Search engines
send 61% of traffic to the average Web site, and referring sites send 17%.
Communications’ site reverses these numbers but still gets most of its readers from
someplace else: 56.6% from referring sites, like Slashdot or Reddit, and 14.3%
from search engines. Google alone sends 13.6%.
Referring sites are Communications’ hitmakers. They sent most readers to
the site’s most popular stories in the past year. Over 75% of the online readers
of “You Don’t Know Jack About Software Maintenance” ( http://cacm.acm.org/
Should We Teach New Software Developers? Why?”
spiked three days after
it was posted. Referring
sites can give stories a
second life as well. “Logic
of Lemmings in Compiler Innovation” (http://
off more than two months
after it was posted.
Referring sites are a Web
version of word of mouth.
The force behind them is
readers like you. Motive
doesn’t matter. Magnanimity, didacticism, and egoism all produce the same results. They put Communications’ articles in front of a larger, mostly different audience. If you think something you’ve read on http://cacm.acm.org deserves extra
attention, the simplest way to share it is with the SHARE button located on the
right column of every article page. The 224 sites accessible from the button at last
count testify to its ease of use. Raising a story’s profile not only brings more traffic to the site, but shares and extends ACM’s profile and membership value to an
even greater, global audience.
WInS ACADem Y AWARD SIGGRAPH Executive Committee Director-at-Large Paul Debevec recently received a Scientific and Engineering Award from the
Academy of Motion Picture Arts
and Sciences. Debevec, with Tim
Hawkins, John Monos, and Mark
Sagar, were recognized for the
design and engineering of the
Light Stage capture devices and
the image-based facial rendering
system developed for character
relighting in motion pictures.
in an email interview,
debevec, who leads the graphics
laboratory at the University of
southern California’s institute
for Creative technologies,
discussed the Light stage
capture devices’ computational
challenges. “our classic Light
stage process built realistic
computer graphics models of
actors by taking photographs
of the actor’s face under
hundreds or thousands of
different lighting directions,
often from multiple viewpoints.
this allowed an image-based
approach to rendering the
actor under complex lighting
environments by computing
linear combinations of the
images taken under the
different conditions. the
imagery was a huge amount
of data, especially when it was
scanned at film resolution.
one observation we made was
that the linear combinations
could be computed even from
image data projected onto a
compressed basis. We could
thus relight the actor’s face
by directly recombining the
compressed image data, and
then decompressing the result.
Using this technique, our Face
demo software (http://www.
relight human faces in real time
even back in the year 2000.”
Currently, Debevec and
colleagues are making the
process of creating animated
digital characters from their
Light Stage data much more
automatic, trying to improve on
the results of their recent Digital
Emily project. —Jack Rosenberger