RE. Text is bad?

["Hart, Geoff" <Geoff-H -at- MTL -dot- FERIC -dot- CA>]

Tim Altom continued to promote the use of graphics:

<<Compared to the age of the Chinese written language, our own "advanced"
alphabet is a kindergarten child.>>

Not in the least. English is far more precise and flexible than Chinese. I
make no pretense of speaking for the Chinese, nor of being expert in the
language, but there's no question that the Chinese system of writing is
cumbersome and inefficient even for native Chinese. Have a look at the
literature on keyboard and OCR development for the Chinese alphabet and
you'll see what I mean. Moreover, while I was in university, several Chinese
friends who had come from China to study in English complain about how
rapidly they were losing their Chinese skills as a result of working in
English. I've never heard of an English or French speaker who worried to
this extent about losing their writing skills while working in another
phonetic language.

<<The fact is that reading, unlike speech, is an acquired skill>>

Both are acquired skills. In fact, there are strong suggestions that the
mechanisms for learning the grammatical structure of spoken language (and
thus, of written language, which closely follows speech) are hardwired into
the brain. Check out Steve Pinker's work, for instance. While it's true that
visual processing is also hardwired, there's overwhelming evidence about how
subjective visuals can be; text is not absolute by any means, but it's far
less abstract, particularly in the hands of a skilled writer. Even skilled
artists never claim to communicate unequivocally.

<<up to 20% of our population is locked out of textual material, and an
unknown larger number is reluctant to use it. A good estimate is that our
text is uncomfortable or unusable for up to 50% of its readers.>>

I have no figures to compare these with, so I won't challenge them. But
whatever the merits of these summary figures, it's far more useful to
consider the context in which you're discussing text vs. graphics. Extreme
example in favor of text: It would be ludicrous to document a grammar
software without using words. Extreme example in favor of graphics: It would
be equally ludicrous to document a color picker utility solely in words.
Just about every real situation falls somewhere between these extremes
(e.g., in neither of my extreme examples would I attempt to use exclusively
words or graphics), and your recommendation that we consider replacing some
text with appropriate graphics is important. As _writers_, we do tend to
forget that words are only one of our tools for communication, and not
always the best tool.

<<Even in our super-literate, whiz-bang, high-techie culture, most people
ask for training from a real-life human, rather than from a textbook.>>

Unfortunately, that has little to do with the ability or desire to read. The
two main reasons why real people make better trainers than books are that
(i) good trainers adapt their approach to the audience and (ii) even bad
trainers can answer unclear or poorly thought-out questions. No book or
online help system yet developed can do either of these things, although
there are primitive steps being taken in that direction (e.g.,
natural-language queries have recently become surprisingly effective... when
the people who wrote the online help covered all the right topics and
indexed them correctly, which is very rare indeed). No database yet written
comes even remotely close to encompassing the total amount of information
stored in a human brain.

<<A surprisingly large number of college graduates are unlikely to open a
book to track down a problem.>>

And that reflects a real, and increasingly important problem for us as
techwhirlers: the MTV generation has been followed by the Internet
generation, and neither generation developed the love for words that we
older folk developed. That reinforces your point that we will increasingly
have to consider nonwritten means of communication. It also returns to your
point that we have to know our audience: the key point is not that text is
ineffective because 50% of the audience has trouble with text, but rather
that we have two audiences: one (the first 50%) needs text and the other
(the second 50%) needs other means of communication. We can't afford to
ignore the first audience by simply saying "text is bad" any more than we
can ignore the second by saying "tough luck; grow up and learn to read".

<<Quality assurance stations now generally use color photos to distinguish
good and bad parts, rather than using descriptions.>>

Certainly. As I noted above, there are some things for which graphics are a
far superior solution. But by no means is this universal. For example:

<<Software demos are done with visual software, not with words.>>

I've yet to see a demo that didn't include either text balloons or a
voiceover to explain what's happening on screen.

<<Text, by its nature, is extremely fallible.>>

Much less so than graphics; there are countless really good word
dictionaries, but damn few useful graphics dictionaries (even visual
dictionaries are extensively annotated with text, and provide multiple text
indexes to help you find the correct graphics). Many people have attempted
to develop means of making graphical communication less subjective (and on a
par with text's relative objectivity), from Jacques Bertin's attempts to
develop a "grammar" for visual communication right up to William Horton's
books on icons and Edward Tufte's work on visual communication. The fact
remains that despite these efforts, nobody has yet made visual communication
as unequivocal as textual communication. Context generally helps to
establish the meaning, at least to some extent, but how do you establish the
context? With words in almost every case. Think of the beach scenes in
_Saving Private Ryan_: is this "brave American boys dying to save the free
world", or "invaders trying to take away our rightful possessions and
meeting a deservedly horrible fate"? Depends on whether you're an American
or WWII-era German, doesn't it?

<<Even the most concrete words are capable of misinterpretation. Even the
word "text" itself has multiple meanings.>>

Yes, but a skilled writer can ensure that any reasonably competent reader
knows exactly which of those words is the correct one. That's considerably
more difficult to do with graphics. I expect that over time, we'll see the
students of Bertin, Horton, Tufte, and others develop increasingly clear and
objective means of communicating graphically. But we're not there yet, and
won't be for a very long time indeed.

<<text is used extensively not because it's an excellent means of
communication, but because other forms are far more expensive. Drawings come
high. Movies higher still. Personal training still higher.>>

Yet all of these are becoming increasingly inexpensive. You can get an iMac
with digital video editing plus a digital camcorder for less than $3000; 10
years ago, you couldn't get the combination for less than an order of
magnitude more. We'll see far more of each of these media used in the
future, but not because they're inevitably more effective. Anecdotal
evidence: Jared Spool gave a lecture a few years back in which he reported
studies of "talking heads" help (in Quicken's online help, I believe). Users
loved the talkies, and rated that help much more highly than conventional
help, yet they also performed considerably less well with the talkies than
when they simply used conventional help. The UIE people are on this list, so
they can undoubtedly provide more details and other examples of this, but
the lesson I took home is that preferences are not always strongly related
to performance.

<<the phone book could easily be done graphically, if the cost wasn't so
high.>>

Not at all. The production would actually be far less expensive if you put
the book on CD-ROM. But ask yourself this: If you were looking for _my_
phone number in a purely graphical phone book, how would you do it? Look
under bearded folks? People who wear glasses? People who live in those
northern 10 states <g> above the Great Lakes? People 6 feet tall? Writers
vs. artists? Corel's clip art collection comes with a book that dramatically
illustrates the problem of trying to create a clear hierarchy for visuals:
why do people come after animals? Because A comes before P in the English
alphabet, not because there's any logical connection between the two. Back
to finding my phone number: I assume you'd look first under males, then
under bearded males, and so on, but it would actually be more logical and
efficient to start looking under beards, then make the second level of the
tree "unbearded: male or female". (That would account for the occasional
bearded woman in your audience, _pace_ John Varley.)

With current technologies, there's no reliable way to organize any large
quantity of information visually; something small like a school yearbook can
work solely because (i) there are few photos (maybe 1000, vs. something like
9 million for the New York phone book), (ii) the people are grouped by class
or teacher or other familiar ordering mechanism (a mechanism that is wholly
absent for New York), and (iii) the people are almost always listed in
alphabetical order, with their names printed beneath the photos. If the
photos are so useful, ask yourself this: Is it easier to refer to me as
"Geoff" or "6-foot bearded bespectacled brownhaired dweller northeast of the
Great Lakes"? How about Alison Burgess, who I had a major crush on in Grade
11? You could find her for me now by searching on her name, but if I asked
you to find her based on my poor attempts at description? Not a chance.

Tim, your basic point is important, and will become increasingly so: text is
only one of many communication tools, and not always the most efficient one,
and we're going to have to grow increasingly skilled in using the
alternatives. But by taking an extreme position ("text is bad") to stimulate
discussion on this topic, you run the risk of going beyond simply getting
people to think and begin to undermine your point.

--Geoff Hart, FERIC, Pointe-Claire, Quebec
geoff-h -at- mtl -dot- feric -dot- ca

Hofstadter's Law: The time and effort required to complete a project are
always more than you expect, even when you take into account Hofstadter's
Law.