Quality of Information, GopherCon94 (April 22, 1994)
Anders Gillner, KTHnoc/SUNET
"The mind is built on an enormous selection of information and it does not
get more knowledgeable because of the information it contains, but because
of the information it does NOT contain....."
Tor N=F6rretranders
"Information should not be mixed up with meaning..."
Warren Weaver
Background
Making things clear from the beginning, this talk is NOT going to be
tecnical stuff... After having dealt with information in the networked
world for a couple of years now, I have found that we pay a lot of attention
to technicalities, not only when we are dealing with networking, but when we
are dealing with information and ordering information as well. I t seems like
some people are satisfied with the mere situation that the information is
there, and is ordered and available, and in my opinion, it is not. It just
might be a good idea to stop for a while and think about why we need all the
information out there on the net, and what kind of results we want from
people working with that information. I spend a lot of time, travelling
inside and outside my country, giving talks to people about the Internet, and
what impact using the net can have on their business or their academic
professional lives, and I get all kinds of questions. A typical question is;
"But isn`t it real cheap to manufacture a CD-ROM today ?" The obvious answer
is "Yes, it is very cheap, but what do you want to do with it ?". At that
time the person asking me the question, looks at me as if I am some kind of
geek from outer space, because for them everything is set and ready when you
got your data down on the CD. Again, obviously that is not so, to make a CD
and fill it with pictures of a certain kind, or text information covering a
certain subject, will cost about a hundred dollars, to turn the CD into some
kind of learning tool will probably cost you around =A4150.000, and I think
that the difference between the stuffed CD, and the learning tool is an
equivalent to the network information services we have today, and what we
could have if we really used the network in the ways it could be used.
The situation today
In an article in a Swedish newspaper recently, a Swedish physicist
said " The global "village"is a big first time experiment". What she meant
was that the we don`t have a "global mind". Since beginning of history, man
has been a more or less geographically static animal. A hundred years ago,
most people did not travel more than a few kilometers a day, and their lives
would have seem utterly boring to us. Today "Netlanders" travel all over the
world in a virtual sense of the word. That article as a whole deals with the
human helplessness in a situation where she has no overview and where no
respect whatsoever is payed to the human way of subconsciously sorting out
information, a situation which I think describes the situation for a lot of
the new users we have , or are about to have, on the networks today. I
sometimes have the feeling that we put users on the top of a slalom pist,
telling them to learn the skill of turning, and the rules of the game on
their way down ! We, "the Netlanders", know the rules, we (occasionally),
have the subconscious ability to sort out the information we want, because
this is where we "live". They don't. We have, in a way passed the threshold
where we do not use only our concious methods to work on the net, we have
started to develop an instinct for solving problems on the net, and what is
an instinct if not just an unconcious feeling for what is right to do in a
certain situation ? The concept of the "Global Village" is not really
something that netpeople use, it is an expression that they use as a
description of the net, in trying to describe their world to the outsiders.
And I am quite sure that a lot of the people on the net really think that
that concept only covers the half of it. New "Netlanders" have to be helped
out, and what we should discuss more is how to do that, or how we could
disseminate the knowledge we have to them.
The Network Hardware Store
So the today 's situation on the net is quite confusing for the
unexperienced user, and we have already seen a couple of solutions which
tries to solve that problem. Almost since the first gophers came upon the
net, we have had "Subject Tree:s" trying to order information in categories
taken from one or another Library Classification Scheme, after the model
"L2=3DLong Nails, L23=3DLong Nails, Blue, etc...". This is of course better
than chaos, but do we really use this new toy of ours in the best possible
way? The normal way for humans entering something new, is to make it look
like something that has been around for a long time , our first cars looked
like horse powered chariots for example. Ok, it is hard for us to take
advantage of computers and networks, but is the solution really to make
computerprograms and network resources look as familiar as possible ? I mean
the kind of thinking that makes a designer design a program logo and
interface for a register program so it looks like a card register ? I
personally sometimes have the feeling that some people thinks "Hey, this new
thing is scary, let`s make it look familiar so people don`t go off their
rocker when they see it !" But the real problem is not what it looks like,
but how it is percieved. If you make people use a digitized card register,
there is a risk that you lock their conceptualisation of the tool they are
using, setting up limitations, not only about possible ways of using that new
tool and the network as a whole, but in human perception of theses new
possibilities, so we are really dealing with psychology here, and we might
really need totally new interfaces to information than the ones we have
today, not just to be able to use the information in the best possible way,
but also so that we communicate to the users that this is something entirely
new
I think that what we need is research in the area. How do humans scan
information ? How do we store information ? We know that the brain does a
primary selection of information more or less on receptor level, and that
information is sorted out very early in the perception process. Let me give
an example, A playground with fifty kids playing, mine gets hurt and start
crying. For me the background noise more or less disappears, because all I
hear is my kid crying. The stronger the feeling is that is associated with a
sound, the stronger is my reaction is when it comes to filter out the
information that is crucial to me, or maybe to my survival. These functions
are old ones, because man is an old animal, and we are in no way designed to
work in an environment like the one we have on the network today. The
information flow today is more likely to bring out another old mechanism in
us, habituation. How many of you have not had, what I should like to call the
"Bosnia experience" ? Day after day, they send out pictures of houses getting
shot and bombed to pieces, people crying, thousands and thousands of refugees
etc. and after a couple of weeks your feelings about the whole thing is more
like boredom than something else, you have had too much similar information,
and another crying child (which is not your own) does not bring up the same
feelings in you that the first one you saw did. You have gotten yourself
habituated... So, we tend to fall back on our old knowledge of the world when
we sort out information, and we also tend to sort out information that fits
into the the knowledge or information that we have before, we , sort of, need
a hanger to hang our coat on, and if that is not there, it takes a great deal
of learning and restructuring of the mind to "manufacture" a new one.
Sorting out Information, How to improve on quality ?
Craig Summerhill from CNI, sometimes refers to the present level of
networked information as the Paleolitic level, and I think that it is
important to bear in mind that we have just started. It willtake us a long
time to get where we want, and the goal will shift during our travel there.
Probably even our way to percieve information as such will change, because it
is quite traditional in it's nature. A lot of the quality information flow
has traditionally been handled by librarians, and the information experts
will probably be with us in the years to come, even though their roles will
change a lot. Instead of sorting electronic books into electronic shelves,
their role might be to determine which expert system that might benefit from
being exposed to a certain piece of information, and even though the system
itself determines if the information really fits, the percental weights of
the information quality has to be determined by the librarian.
In the future we will probably have networking tools that do that
primary sorting for us. We are talking about years of experience here, maybe
it is so that "growing up" in a networking sense means that you use years
instead of hours to set up the filtering mechanisms that you need to be an
experienced network information user. We have to use computers and networking
as extensions to our primitive selves, not as tools as primitive as we are .
When I worked as a psychologist during the 70's, someone at
psychology department at the University of Gothenburg came up with a
tachistoscopic test that was said to test man's psychological defense
mechanisms. That might be dubious to some of you, but the fact is that when
they tested combat pilots in the Swedish Air Force, they found that the
result predicted with 90% accuracy which of these who was going to have a
crash during the next couple of years. What they tested was a persons ability
to percieve a picture without shunting out some of the information in it. It
seemed like some persons consistently threw away information that was
threatening to them, and obviously, these did not grow personalities that
made them good pilots... Humans tend to act both psychologically and
physiologically during stress ,which this is a good example of, and we are
not only shunting away information that we don't like, we also tend to, like
in the crying child example, to focus on information that we like, our pupils
dilate when we look at somebody we like, our pules rate goes up when we fall
in love or see something that we really like, and we also get GSR reactions
when we hear something that is good or bad, and these reactions does not
neccesarily, have causes consciously known to us.
What I want to say with this is that we really don't know much about
our "own" idea of the world, the unconscious sometimes makes choices for us
that we are not aware of, and sometimes the choice is made on a very
primitive level, fooling us to think that we act optimal in a situation when
we don't do that at all, but using more than one of our senses while sorting
information seems to be a good idea
Making it better in a short perspective
This might be the future, but how to improve the quality in a short
perspective ? Let us make good use of the librarians we have, and of the
systems we have. Putting librarians to work with roaming the network for
items to sort will of course not scale in the long run, this will have to be
done in a distributed way. The local librarian should only be responsible for
the classification of the material that her/his site put onto the network.
Use any classification scheme, it does not really matter, because the good
thing is not the scheme, but that the resource is stamped as a quality
resource. If we will be able to make out the difference between quality
checked resources and the rest, we have improved on the information quality
on the network in a terrific way, compared to the situation today, and if we
are able to do searching with a parameter that tells the search program to
look for quality resources only, we have improved it even more. We will still
have the Network Hardware Store, but we will at least be sure that some of
the boxes will contain nails with a certain reliability.
There is, of course, also the possibility to make more than two
categories of information, separating out for example news, university
courses, K12 information and adminastrative information will make it even
better. So we need a lot of librarians out there, and who will pay them ?
Information is costly, and knowledge will cost a lot more still.....
Information reduction is always costly, and as in science, data
reduction will cost you some of the valuable content as well while reducing
the noise that you really wanted to get rid of
The part you really want to get rid of is what sometimes is called
"exformation", and to get rid of that will always cost you, knowledge, time,
money, you name it. Somebody will always pay! There has been a tendency on
the net, to argue for a situation where all the information should be free,
and that is possible when you put your own free time onto deriving
information out of data for the benefit of the net population. "How to make a
japanese kite", or "The Complete List of Beatles Songs" might be free on the
net in the future, but I do not think that the Reuter statistics, or the Jane
Fonda Workout program will ever be free. Making this information (or whatever
you like to call it:-) )has cost some commercial company big bucks, and they
are not likely to give it away for free. We are about to get commercial
information on the Net on a big scale, and we better be prepared for ....
Coca-Cola Information.....
Netland has it's advantages compared to other countries. The average
level of intelligence , for example, must be very high, and that situation
will not change as long as the net is, more or less, academic. Net society
today share a lot of views about the world, we are all more or less , well
eduacated, which among other things, means that we share a certain view about
our environment, about the neccesity of research, about the neccesity of
education, about the neccesity of beeing aware of the bias of our own
perception, etc, but what will happen when systems like Minitel brings other
groups onto the net ? The Net will have some inpact on these people, but they
will have some inpact on the net community as well. As soon as we open
Pandora's box, we will have all the problems that other media have and have
had during the past 400 years on our back, and if we don't think about policy
before that, everybody that thinks that their ideas are the right ideas and
that nobody else's ideas have the right to live, will try to control what
others do and have available. So we are in for a change, and what can we do,
to make that change an easier one ? Not very much that I could think of right
now, but being aware of the storm at least give you the possibility to
prepare for it. The market could be both a problem, and an advantage. The
market will sell us information, but we should also be concerned about too
much concentration of power around any special player on the information
market, just as such concerns has risen when a similar situation has become
reality in on other market. He who controls information controls the world is
a saying that I don't think that anyone of us wants to see become a reality,
although we want it to come true in a broader perspective. The authors of an
article about what they call "cyberwar " say that information will be the
next crucial point when winning the next war, as it had not always been like
that....
The filtering of information has always been something that censors and
others have used as a tool in warfare, and I really think that they have a
harder time to filter today, than they had in old times. Using intelligence
information has become easier though. What we don't want to have on the Net
is a totally controlled information market where dictators, companies or
others that like to think that they are controlling what we shall know or not
know, is ruling.
Communicating Information, not as easy as it might seem
Now if I want to communicate something to someone else, how do I do
that ? I speak to him or write it down for him to read. Which means that the
information just has made it through the first filter
Speech is information ordered already, because putting words on
something is a filter in itself, compare somebody's notion of a horse, and
the word horse, which contains only a fraction of what the person saying it,
means by "horse", which can be totally different from what the person
listening means by "horse". Human communication really has its drawbacks, and
the big question is if the network , and the new technology could make it any
better. Man has always tried to write laws that were meant to regulate how
man should live together with his peers in society, and just see were we have
ended up. The language that we use is really an extremely inexact
representation of what we really mean, in spite of body language and other
non-verbal ques. Using other types of communication on the network, like we
do in real life but in an extended way, and using all our senses, might give
us an opportunity to improve on information quality in centuries to come.
What do we want the information to do for us ?
What we really need is not information, but knowledge. We need people
to expand their knowledge of the world, and we want our kids to be able to be
better citizens in the future. So how do we get knowledge out of a system
that only offers information ? Well, to get knowledge you need knowledge, so
we are heading into psychology again, but learning psychology this time.
There is really no learning psychology designed for network use, although
there are quite a few attempts, mostly associated with new CD-ROM learning
tools, and some of these will of course work on the network as well, but to
be able to find your way through an encyclopedia, you need to know how to do
that , so we probably have to improve the basic skills of reading if we want
this to expand outside the educational and commercial sector. If you take a
look at a network map today, you will find that networking is an industrial
world phenomenon, and that there must be a marvelous market for distance
education in the developing world. I can see networking and distance
education as one of the main streams of foreign aid in the coming years if
the institutions giving that aid are smart enough to see the potential it.
Helping the developing countries out in this respect should be one of our
prime concerns for the future, if we don't want to carry on in a world of
have's and have not's. But we do need research in the learning psychology
field to make the most out of networked information, and turning that
information into valuable knowledge, might be the next great challenge for
us.