Rise
of the information marketplace
The
sharing of information and e-Commerce over today's
Internet is only the tip of the iceberg.
"Content"-- TV, movies, radio, books,
magazines, etc.-- accounts for less than 5% of the
world's economy. However, 50% of the economy is
information (office) work, which is barely happening
over the Internet today.
Integrate
computers into our lives
What's
all this talk about "getting on the
Internet", like it's a special place you go?
Today, we are at the beginning of the information
revolution. The last revolution was the
industrial revolution. Its greatest innovation
was the motor. Did we have special places just
for motors? No--we brought the motors into
our lives, integrating them to such an extent that we
no longer mention the motors, only the tools in which
they are embedded. The same thing should happen
with computers. Don't tell me about its
specifications---tell me what it can do.
Give
us a gas pedal and a steering wheel
We
need to raise the level of controls we use with our
computers from the current low machine level to a
higher human level. Imagine if on your car you
had to individually control spark plugs, fuel mixture,
brake tension, etc. No! In cars, we simply
have a gas pedal, brakes, and a steering wheel.
We need the equivalent for our computers.
Reach
all people
Only
5% of the world's population is connected to the
Internet. How can the web be worldwide when it
only reaches such a tiny portion of humankind?
If this is allowed to continue, it will only increase
the gap between rich and poor. Instead of
providing more training, why don't we make the
computers easier to use?
2
LET'S TALK
"Our
computers are hard to use. They enslave us
rather than serve us."
We
should throw away 90% of the features in today's
bloated software.
You
were born with eyes, ears, a mouth: the ability
to see, hear, and speak. You were not
born with a keyboard or mouse socket. Human
centric computing says you should use these natural
human abilities to communicate with machines.
Elusive
intelligence
There's
a lot of hype about "intelligent agents",
but current ones exhibit only a sliver of
intelligence.
Speech
and vision: different roles
Humans
use speech for two way communication. But vision
is used mostly one way-- for taking in information.
therefore, speech should be the primary means of
exchange between people and machines, while vision to
be the primary approach for getting information from
the machine.
Let's
talk
Speech
recognition systems are difficult. Speech
understanding programs, in comparison, are easy---they
require no training and can understand anyone.
We need to shift power efforts towards speech
understanding programs (examples from LCS research).
Speech
readers are getting much better; they sound less
computerized and more natural.
There
needs to be growth in translation services as commerce
becomes more global.
Show
me
Display
from machine to the human is progressing: bigger
displays are in the works, as are three dimensional
displays. Head mounted displays, VR displays and
augmented reality displays are also progressing.
However, visual recognition from you to the machine is
lagging. new research in visual recognition,
intelligent rooms, and pattern classification may help
us there.
A
new metaphor
Are
two main metaphors today are the desktop and the
browser (window shopping). Each is good, but
together they make no sense because they work in
different ways--- it's like requiring different
dialing systems to make local versus long distance
phone calls. we need to merge these two
metaphors into a single system. One possible
metaphor is that of servant. Another is a
virtual geographic metaphor (Bob: kind of like William
Gibson's "cyberspace")--- moving in space is
a natural human endeavor with which we have thousands
of years of experience. Similarly, a "historicoptor"
would allow you to drill down into the past by
pressing down on a joystick or into the future by
pulling up.
Brain
chips
Direct
human/machine interfaces are tempting, but we would
have to deal with the problem of information overload
before brain implants are a good idea.
3
DO IT FOR ME
The
author uses an "e-bulldozer" to book his
airline travel. a command given in a few seconds
produces about thirteen minutes of work by the
computer that replaces perhaps hours of work by a
human, similar to the way a slight touch on the levers
of bulldozer can lift thousands of pounds. This
automation is one of the most promising aspects of
human centric computing. Properly applied, such
automation should be able to raise human productivity
by 300% during the next century. Since, as noted
earlier, office work accounts for more than half of
the world's economy, this is a big deal (you could do
a year's worth of office work in four months!).
The
ascent to meaning: e-forms
Computers
should be able to "understand" what they are
communicating to one another so they can act on it.
For example, computers should be able to automatically
fill in forms for us. Failing that, forms should
be speech driven, underlain by speech understanding
programs.
Problem:
how do we agree on systems of shared meaning?
Perhaps by either gradual adoption of standards or
through "Semantic Webs". Metadata
(e.g., XML) shows great promise.
Bring
things under control
Automation
should extend to control and coordination of physical
devices around us.
Hundreds
of dumb servants
A
bunch of little "scripts" could automate
daily activities. Start
small---use "Quickkeys" or other macro
programs to automate common tasks.
Automation
and society
Automation
will result in the loss of some jobs, but mostly jobs
that are repetitive and require little human
intelligence.
4
GET ME WHAT I WANT
Individualized
information access.
"Today's
information retrieval systems don't understand what we
mean what we ask for something... and they don't
understand what all the information they sieve through
is about."
Organize
or search?
How
do people get at information without computers?
We check our desk drawers, we ask friends and family,
we look in reference books and contact experts.
Our machines should do likewise---check themselves
first, then go to machines of our friends, and roam
the information marketplace.
We
need a "meaning processors". Meaning
processors can be "extractors" that pull
header information from a file, or
"observers" that track the frequency with
which you use each piece of information you touch, as
well as linkages among the pieces of information you
use. These media processors would work
automatically, all the time.
The
semantic web conspiracy
Tim
Burners-Lee, the inventor of the world wide web,
dreams that the web will gradually form a gigantic
"brain" by adding the capability that can
relate the meaning (or semantics) of the information
contained in web pages. Manufacturers can aid
this process by joining together into groups in
deciding on standards for using XML or RDF to code
their web pages.
A
new information model
The
prevailing model of information storage on our
computers right now is that of hierarchical files and
folders. Meaning is not included in this picture
(except maybe for file names).
The
web is not hierarchical. Imagine if all of the
information on your computer was linked together using
hyperlinks, like on the web. Pictures, audio
files, video files, and text files related to a given
concept would all be automatically linked to each
other. Now imagine extending that to the
computers of your friends and associates, or even all
the computers on the web.
Call
to action
Ask
yourself the question: what information is
reliable, timely, and vital to me or my organization's
purposes? It's in your best interests to find
the best information out there, and that information
is not standing still. You can start by
searching for good information sources. Ask
peers, look at competitors, and search the web.
The harder you work to find that information sources,
the easier finding information will be.
Another
big question to ask: what information within my
organization is not currently on any machine but could
be dramatically more helpful if it were computerized?
Pick areas where the potential payoff is great.
5
HELP US WORK TOGETHER
The
challenge
Now,
via the web, we can electronically reach any one of a
few hundred million people around the world-- that a
thousand times more people than we could reach with an
automobile, and a million times more people than we
could reach on foot. Our challenge is to convert
this proximity into useful person-to-person and
organization-to-organization collaboration.
E-commerce
brings buyers and sellers together; however, hardly
any of these transactions involve direct negotiations
between buyer and seller, so this is not really
collaboration.
Messages
and packages
Currently,
the most widespread collaboration is via email (with
an estimated 1.5 billion messages per day).
Future email will include more speech, diagrams, and
video.
However,
email can sap your time. We are headed for an
estimated ten fold increase it received messages, so
much that it is estimated that within ten years your
email will require eight hours of daily attention.
Clearly, something must be done.
What
to do? "Birth control at the source, and
euthanasia at the destination".
Birth
control: consider---if you send an email message
to one other people, you'll consume half a day of
their collective lives!
Use
filters to remove unsolicited email and categorize
other mail. use macro programs to automate
common email tasks.
Collaboration
systems
Tomorrow's
collaboration systems need to provide us with the
natural feel of face-to-face encounters.
Consider the telephone. It's "low
tech". However, it's successful because it
approximates the naturalness of a spoken conversation
held face-to-face.
To
do this, we need three things:
1
making distant encounters realistic, as if they were
held in the same location
2
carrying forward the "meanings" pivotal to
continuity in asynchronous encounters
3
coordinating person to person encounters with the
other human centered technologies people we use with
their machines.
Currently,
delays and glitches more this process. But this
will change. Likewise, video is currently small and
grainy, but it will get better as bandwidth increases.
Sharing applications is already easy, but controlling
them at a distance is still difficult.
For
asynchronous communication, we need collaboration
editors that will do for the collaboration process
what text editors have done for document preparation.
Information
work
Collaborative
software will soon be customized to professionals and
tasks. Medicine, salespeople, real estate, lawyers--
even video games and dating services.
Privacy
Information
technologies can be used to attack on privacy, but
they can be used to protect it, too.
More
social consequences
