快猫短视频

BOB HORN likes to tell a story about a scientific meeting he attended a few
years ago. 鈥淎 biologist got up to talk and the projector failed,鈥� he says. 鈥淪he
said, `I can鈥檛 give my talk without the slides.鈥� A thousand people in the
audience groaned. I was the only person who was excited.鈥�

No, Horn isn鈥檛 a connoisseur of embarrassing moments. He was excited because
the speaker鈥檚 misfortune was a perfect, though unplanned, example of how words
and images can become so tightly interwoven that neither can stand alone.
Indeed, Horn, a visiting researcher at Stanford University near San Francisco,
thinks text and pictures have become so interdependent that they have become a
language in their own right.

Horn believes this 鈥渧isual language鈥� will change the world. He says that it
is far more efficient at conveying complex ideas than conventional methods of
communication. Armed with the ability to read and write visual language, we will
be better equipped to cope with the 鈥渇irehoses of data鈥� that threaten to
overwhelm us today. Its big advantage is that it allows people to digest in
hours complex issues that might otherwise take weeks to understand.

But most exciting of all, visual language is a tool for mapping ideas. Horn
and others have used it to map out complex manufacturing projects, to chart
political arguments and to navigate the treacherous waters of scientific and
philosophical debates. Their charts allow the viewer to quickly grasp the main
points of a debate and its current status. They even show the edge of
knowledge鈥攁 feat that is impossible by other means.

鈥淲e鈥檙e preoccupied by information overload, but we forget that information
overload is less a function of the volume of information washing past our eyes
than the gap between the volume of information and our capacity for making sense
of it,鈥� says Paul Saffo, a technology forecaster who heads the Institute for the
Future in Menlo Park, California. 鈥淏ob Horn is going after big game because, if
his theories work, they will help us make our sense-making tools catch up with
the volume of information washing over us.鈥�

The idea of combining words and images is hardly new, of course. Humans have
drawn pictures鈥攁nd presumably talked about them鈥攆or tens of
thousands of years, far longer than they鈥檝e known how to read or write. And over
the past two centuries statisticians, political scientists and business
theorists have developed a large repertoire of graphs, flow charts, timelines
and other methods to summarise complex data and procedures. But three changes
have combined to thrust visual language onto centre stage in recent years, says
Horn.

First, modern life has become more complex. People in fields as diverse as
manufacturing and science have a bigger pile of essential information to
assimilate and less time to do it in than their parents or grandparents. That
makes it crucial to provide efficient overviews that let people pick and choose
areas where they want more detail. Second, with computer graphics now the norm,
almost anyone can merge text and images into a polished-looking document, no
matter how clumsy they may be with an artist鈥檚 pencil. And last, the video
culture has given people, especially the young, much more visual sophistication
and made them more receptive to receiving information in more than just
words.

鈥淰isual language is being invented all around the world by engineers and
consultants and mathematicians. It鈥檚 being invented out of need, to express
complex ideas,鈥� says Horn. The past two decades have already seen the emergence
of a new cadre of content-oriented graphic designers, sometimes called
鈥渋nformation architects鈥�, that specialise in finding the best combination of
graphics and words to convey information鈥攆or example, in the explanatory
graphics that now appear in many newspapers and newsmagazines.

But visual language is much more than just good graphic design, claims
Horn鈥攊t is a true language, complete with its own formal rules of syntax
and semantics. Most people already know many of these rules from years of
experience with visual media such as cartoons, television and the adverts in
magazines. We know to read down lists, that arrows can show the flow of time or
ideas, and that a series of concentric rings can represent a target. We
recognise the conventions cartoonists use to indicate motion, collisions, and
emotions. In short, almost everyone already knows how to read visual language,
says Horn. All they need is to learn how to write it well.

Horn may go too far in claiming that this marriage of words and images is a
true language, says Talbot Taylor, a linguist at the College of William and Mary
in Williamsburg, Virginia. In real languages, rigid rules of syntax make certain
arrangements of words鈥攕uch as 鈥淎te supper my I鈥濃�攚rong and others
right.

By contrast, the laxer rules of visual language say only that some usages
work better than others鈥攁 list that starts in the bottom right-hand corner
of a page is not as easy to read as a list that starts in the top left-hand
corner. However, says Taylor, that doesn鈥檛 detract from the power of visual
language. In fact, he says, the freedom from rigid rules is a good thing. 鈥淚
think it鈥檚 a mistake to try to explain the creative power it has in terms of the
narrowly restricted communication power that language has,鈥� he says.

There鈥檚 no doubt that images can help text get its message across. 鈥淭here鈥檚 a
lot of power in integrating words and pictures. It really can improve people鈥檚
understanding,鈥� says Richard Mayer, a cognitive psychologist at the University
of California at Santa Barbara. A few years ago, for example, Mayer tested how
well volunteers could understand the workings of a bicycle pump or a car鈥檚
brakes from verbal descriptions alone, pictures alone, or both simultaneously.
People could recall details equally well from any of the three types of
presentation, he found. But when the volunteers had to apply their
knowledge鈥攂y answering questions about how to fix a broken pump, for
example鈥攖hose given both words and pictures came up with about 50 per cent
more correct answers than the others.

That boost probably comes because the brain uses different pathways to
process verbal and nonverbal information, says Mayer. A person who gets
information from both channels at once has an easier time linking up the pieces
into a meaningful whole.

Because of this, visual language works better for some kinds of information
than for others. 鈥淚t鈥檚 best at being able to grasp things in context and see how
they鈥檙e related,鈥� says Terry Winograd, a computer scientist who directs the
Program on People, Computers, and Design at Stanford University. 鈥淚t鈥檚
correspondingly less good at precision and detail.鈥�

Indeed, for many overview tasks, visual language is already so commonplace,
it鈥檚 unremarkable. 鈥淲hen you sit down in your car, you don鈥檛 read a bunch of
text. In the cockpit of a jet, you鈥檙e reading enormous amounts of data, but it鈥檚
all visual,鈥� says Eric Oksendahl, who directs the design of instructional
materials for the Boeing Aircraft Company in Seattle.

One of the big advantages of such an overview is that it allows people to see
a complex subject in context and focus on whatever bit might be of interest at
the time. 鈥淲e just don鈥檛 have time to read everything,鈥� says Horn. 鈥淚 think
that鈥檚 an important principle for communications in the 21st century. People
should have the opportunity to scan and skip.鈥�

The strongest example of this is an overview technique Horn calls
鈥渁rgumentation mapping鈥濃�攖he charting of ideas. In this, mappers gather
together all the arguments on a given question, together with their rebuttals,
counter-rebuttals, and so on, and lay them out on a giant flow chart. Anyone
looking at the map can see at a glance which branches of the argument have
spawned the hottest debate. And by looking more closely at the terminal twigs of
a branch, one can see who has had the last word so far and what they said.

Thorny question

As an example of argumentation mapping, Horn and his colleagues have tackled
one of the thorniest questions in the field of artificial intelligence: can
computers think (see chart)? 鈥淢y demand was that we do it with
an industrial-strength example, not some little toy,鈥� says Horn. He started with
mathematician Alan Turing鈥檚 bold claim in 1950 that machines will be able to
think someday. Then he and his mappers鈥攖hree graduate students in
philosophy鈥攄ived into journals and books to trace the fierce arguments
that followed.

In the end, their project turned out to be anything but little. Horn
estimates that his team laboured for 7000 hours over five years to create the
map, which runs to seven wall charts over a metre wide and catalogues over 800
individual points raised by 400 scholars. 鈥淚t took a couple of years to perfect
a smooth read of these maps,鈥� he says. The next project (Horn has a grant to map
鈥淲hat is consciousness?鈥�) should go much quicker, he says, now his team knows
how to go about their task.

Along the way, the mappers discovered another benefit of the process: that
the issues became clearer as they charted them. 鈥淭he discipline that this kind
of summarisation and diagramming provides makes the differences [between
positions] emerge more clearly. In prose it鈥檚 very easy to fudge or be slightly
ambiguous. It鈥檚 very difficult to get away with that in this kind of
methodology,鈥� says Horn.

In practice, the charts provide a good example of both the strengths and
weaknesses of mapping ideas. 鈥淥n the one hand, the maps simplify the positions,
so they鈥檙e not as effective in giving you the nuances,鈥� says Peter Suber, a
philosopher at Earlham College in Richmond, Indiana, who uses them in his class
on artificial intelligence. 鈥淏ut on the other hand, articles and books can鈥檛
convey all the connections with the same clarity and speed.鈥�

Indeed, Suber thinks similar maps may be useful far beyond the arid halls of
academe, in the decidedly steamier thickets of politics. 鈥淜osovo is a good
example of a topic that has probably lost a lot of the public,鈥� he says. 鈥淎 map
of this sort would show people what the disputes are. The problem would be
getting the maps in front of the people who could do something.鈥�

Others are already planning to use the maps to guide policy. In Calgary,
Alberta, a group of futurists called the Capitalizing on Change Project will
spend the next year or so developing a map of the argument over the role of
carbon dioxide in global climate change. 鈥淭hat argument has lots of ideologues
and often not much clarity,鈥� says Ruben Nelson, who heads the project. 鈥淲e鈥檙e
hoping that by developing an argumentation analysis we can help people stop
hyperventilating and get clear on what things they disagree on, and which of
those are factual and which are emotional issues.鈥�

The best thinkers of every generation have been able to keep such an overview
in their heads, Nelson thinks. Putting that overview into a concrete form could
give more people the same advantage and bring more voices into social
debates.

鈥淚 call it the language of complexity,鈥� says Elsa Porter, a management
consultant in Portland, Oregon, and a former assistant secretary of commerce in
the US government. 鈥淚t鈥檚 the first time we鈥檝e had a grammar and a syntax that
allows you to speak the language of complexity in a way that ordinary people can
understand. Eventually, I imagine, you could have a conference on an issue, say
reduction of teenage pregnancy, and as people talk about all the causes and
effects, you could use the computer to put a map on the wall. Then people would
really see and understand the causes and effects and what to do about it.鈥�

There are a few hurdles to clear before Porter鈥檚 dream becomes a reality,
however. For starters, anyone without bucketloads of money will have a hard time
even getting hold of a high-resolution computer display the size of a wall, let
alone a 鈥渓ive鈥� one that records what they write on the display. The
laptop-driven projection displays that are now commonplace in lecture rooms
merely enlarge a standard screen, which does not give enough resolution to
display more than a corner of a large map. A few companies have begun marketing
projection systems that can process more detailed live images, but they cost
many thousands of dollars. 鈥淗ow soon will we routinely put a blackboard-sized
live screen in our offices at a sensible cost? Probably ten years down,鈥� says
Saffo.

Undoubtedly, the biggest challenge is to find a way of writing visual
language quickly. 鈥淲e have not even begun to build software tools to help us do
this,鈥� says Saffo. Working with text and images together in existing computer
programs usually involves carefully defining text and graphic elements and
positioning them鈥攏ot exactly an effortless process. 鈥淵ou don鈥檛 want to
spend your time planning how to make the information appear. You want it to just
flow,鈥� says Winograd. 鈥淚f you watch somebody who鈥檚 good at [visual language],
they would be better off with something that felt like sketching with a pen than
like placing something with a draw program.鈥�

Writing freehand

Already there are hints of such an approach. Popular programs such as
Powerpoint include tools for producing tree diagrams and organisational charts
quickly. And a program called SKETCH, now under development at Brown University
in Providence, Rhode Island, recognises simple strokes of the pen and fills in
appropriate detail.

If the user draws three edges coming from one corner of a box, for example,
the program will complete the box. Through a series of steps like this, the
program takes a rough freehand sketch and transforms it into a polished-looking
drawing, complete with straight lines, parallel edges, and so on.

But what makes SKETCH so promising is the next step in its processing. If
someone tells the program that it鈥檚 dealing with furniture, the user need only
sketch a cube and the program fills in a fully detailed chair; for a rectangular
box, it would fill in a table. If a different user tells the program that it鈥檚
dealing with computer hardware, it would fill in similar sketches with different
detailed objects. The techniques for doing this are well advanced, says Robert
Zeleznik, the computer scientist who heads the SKETCH team. 鈥淚n many cases, it鈥檚
just a matter of spending the time to flesh out a complete database of
components,鈥� he says. Eventually, they hope to add voice recognition so that the
user can say 鈥淲indsor back鈥� or 鈥淐hippendale鈥� when sketching the cube and thus
specify not only a chair but the exact style.

Zeleznik and his colleagues have not yet tried to apply SKETCH explicitly to
drawing out visual language, though Zeleznik agrees that 鈥渢hat鈥檚 a direction you
could go along鈥�. However, Winograd and other experts think it may prove to be an
excellent starting point.

Even if hardware and software developers can clear all these technical
hurdles, visual language faces another barrier: most people can express what
they need to say much more quickly by writing plain old prose than by
constructing an elaborate information graphic. Even Horn admits that he could
have covered the ground of 鈥淐an computers think?鈥� in far less time if he鈥檇
simply written a book.

Of course, speed isn鈥檛 everything鈥攐ne can write bad prose even more
quickly than good prose, for example. For many uses, such as instruction manuals
or the executive summaries of business and governmental reports, writers may
find that visual language gives them better clarity or greater mass appeal that
is well worth their extra time. A quick glance at any glossy magazine shows that
most advertisers already think so.