|
|
|
Hypertext & Computers:
data, texts,programs
|
Despite the
novelty of hypertext, it comes with cultural heritage: the
culture of the computer as a scientific tool, and the
culture of postmodern literary criticism. Our model of
hypertext comes as much from these two cultures as from any
in built characteristics. This is part of the hypertext
problem.
|
|
|
The notion of hypertext arose within the context of computer
communication: data-based, complexly interrelated, and fast,
as we operate the machine not to produce a physical output,
but to read its states as it processes. Hypertext is a
natural and intuitive extension and evolution out of
computer information codes &endash; particularly ascii
&endash; and of computer databases. From this, hypertext as
a notion gets three of its major characteristics:
1) its bias toward text (letters and numbers),
2) its reuse or recombination of text and
3) navigation of text data by pointers or
links.
|
|
|
How is text different from image? In Western culture, there
is a very large literature discussing the relationships
between text and image and a historical tendency to see the
two as distinct and often opposing forms.2 Within this literature, text
and image are often opposed on the basis of our greater
ability to parse text grammatically, and thus to pin down
its meaning. Nelson Goodman in particular, has distinguished
text from image by its single system coding, i.e., language,
and our agreements on the meanings of words. In images, by
contrast, there are often not only multiple readings, but
there are not necessarily any set rules for deciding which
aspects of the image are to be considered. Thus, if it is a
painting, we may or may not consider the brush strokes or
the texture of the canvas as relevant to the content. If
that is the case, the text can be rendered by any
duplication of the symbols on any sort of medium, while the
only reproduction of an image is the image
itself.3
|
|
|
By and large, computers operate on images without parsing
them; e.g., copying them pixel by pixel, by region, etc.,
without hierarchy or relation beyond belonging in the set or
file. (Images can be coded for example in medical data or
satellite gathered images.) In contrast, within text, we
presume coding. Computer languages are built using grammars
of textual and numerical symbols which can be analyzed in
the same way we solve mathematical formulae. Diagrams and
formulae actually operate in a middle ground. Formulae look
like language but operate via graphical transpositions
&emdash; move a divisor across the equal sign and it becomes
a multiplier. Diagrams look like pictures, but language-like
rules govern their interpretation. Thus, we often see the
rendering of numerical data and propositions in diagrams as
literal: a diagram can be correct or incorrect, and two
different looking diagrams can be identical in the
information they convey. Similarly, an image can generally
agree or disagree with a description: this person has two
eyes. But no two different pictures can be deemed logically
identical, and in general, the translation from language
(e.g., story) to image is an "adaptation," or
"interpretation," which is to say a translation into a
different medium in which there are no precise
correspondences, only certain overlaps.
|
|
|
What distinguishes text is not only the set data that is
deemed relevant, but also that the rules for interpreting
make it possible to disassemble the data, into sentences,
phrases, clauses and words &endash; with presumably discrete
meanings &endash; and then to reassemble it again. These
ideal types: text as precise or "digital," and image as
"analogue" are often the standards by which we form our
expectations, and they establish the validity and
credibility of images and texts.
|
|
Hypertext
& computers Text conditions image
|
One interesting wrinkle is that markup languages like html,
the language used on the internet, employ coded instructions
governing appearance. These provide a computer-related
approach for looking at visual effects in language-like
terms, offering a possible route to building visual
taxonomies. Computers add yet more layers between setting
and communicative essence like screen size, rendition in
greyscale or color. Creating for the computer includes
designing for a variety of different output devices and
readers all of which will look different to the user, but
all of which are expected to present the same information to
that user. The computer makes obvious what has always been
true. "Éthe central point about electronic textuality
&emdash; its fundamental distinction from the object on
which it is read."4
|
|
Hypertext
& computers Computer
information
|
We are more likely to think of computer text as impersonal
information. The symbols computers manipulate are often
machine-generated: data, shaped by programs and produced as
output. Even our language biases us toward seeing computer
data as objective, non-human and certainly not as human
narrative. We tend to see such output as information that
appears independent of author and of receiver &emdash;
something discovered. Email, news groups and web pages may
be reintroducing authorship and narrative, but they do so
within a space of computer information that did not exist
until very recently.
|
|
|
We can see this bias toward text, toward impersonal writing
and reading environments and toward a belief in the
universality of knowledge in the writings, including this
essay on computer style sheets:
Since the beginning of the Web
just a few short years ago, html designers have had to serve
two audiences. On the one hand, the prime audience for web
pages is, rather obviously, people. Yet the html language
syntax is geared more towards creating machine-readable text
rather than text meant to look good for readers. For
example, there are html commands for declaring that a text
chunk is a definition, an ordered list, or a table; there
aren't commands for setting margins or font.5
|
|
|
Here is the
search for universal knowledge: Computers and hypertext were
often seen as ways of building universal objective knowledge
systems. Early architects of Hypertext theory saw hypertext
as providing a method for constructing a universal
environment for human information use, e.g., Ted Nelson's
Xanadu:Éthe Xanadu system provides a universal data structure to
which all other data may be mapped É (Nelson, 1987, p.1) But the
developers have much more in mind than a computer data
structure. They see writers and readers throughout as
working in the same conceptual space.6
|
|
|
Let us stop here to ask a question "What is the relationship
between data and knowledge?" We think of knowledge as being
something closer to idea, or concept. How would we order
such terms as data, information and knowledge, and let us
add wisdom. We could order them on a line: data is stuff,
information is data selected and arranged according to some
principle or goal, knowledge is the principle or goal known
through the information. Here, we move from the most
concrete to greater levels of abstraction, into which lower
levels fold. This is the process by which perception becomes
cognition, and by which events are glued into larger
structures. When we come to wisdom, we have lost direct
touch with the most concrete levels. Thus, we have two
poles: concrete and abstract which we traverse up by concept
formation and back by verification. The uses of terms like
information and data elide or disregard this characteristic.
They beg the question of whether computers can think by
presenting it as a non-question.
|
|
|
In the 1945 Atlantic article "As We May Think," Vanevar
Busch was concerned with the explosive growth of research
and information, and within it, our inability to process as
its limitation. His solution to the information glut was
greater human efficiency with the aid of the "memex."
...Consider a future device
for individual use, which is a sort of mechanized private
file and library. It needs a name, and to coin one at
random, "memex'" will do. A memex is a device in which an
individual stores all his books, records, and
communications, and which is mechanized so that it may be
consulted with exceeding speed and
flexibility...7
|
|
|
Thus,
within at least that early computer information model we see
tendencies to look at text as autonomous information, and
the implicit view that packets of information have an
integrity like standardized components used to make novel
entities rather like modular housing. By the same token, a
person is a human information
processor.
|
|
|
The human processor is machine-like, which is to say
logical or rule governed, but it links information by
rules of association. In his article, Busch defines
"association."
- The
human mind does not work that way [alphabetically]. It
operates by association. With one item in its grasp,
it snaps instantly to the next that is suggested by
the association of thoughts, in accordance with some
intricate web of trails carried by the cells of the
brain. It has other characteristics, of course; trails
that are not frequently followed are prone to fade,
items are not fully permanent, memory is transitory.
Yet the speed of action, the intricacy of trails, the
detail of mental pictures, is awe-inspiring beyond all
else in nature.8
Hypertext,
then, is information organizable according to the operative
rules of the human brain: a technical solution of the
perceived need to increase the efficiency of the human
information processor.
|
|
|
The term "association" has two aspects: the lack of a
narrator and the problem of deriving concepts from data.
Hypertext lacks the author who selects, arranges and edits
the text (here the data) to build a structure of
interrelated ideas or concepts, thus making it information.
The crux of the problem, then is how to find or make the
glue that bonds these separate texts together into larger
coherent and intelligible units. In this context, Busch uses
the notion of "association" which has a long history in
Anglo-American philosophy. It argues that we build knowledge
by associating things together. Unfortunately, when we build
breadth of knowledge, we must also build depth; in order to
make sense of all of the data we have collected, we need to
provide a concept that orders it. This barrier has dogged
association theory. By contrast, Gestalt psychologists
emphasize that for us, the concept is a container for the
data and often of a fundamentally different category from
the data. Developmental psychologists, like L.S. Vygotsky
9 come at this problem from two
different directions: that our notions of things become more
differentiated, and that the ability to form ideas is not a
spontaneous process at all but is socially passed from
adults to children. Association theory fails to deal with
these issues. Within that theory, having knowledge grows
more directly out of data possession rather than concept
formation.
|
|
|
Another aspect is the distinction between observation and
message reception. Within the computer, information exists
both as data to be interpreted and as messages or
programming. First, information is the universal stuff of
which everything is made. Without information (or data),
there is nothing to work on. But within computers,
information has another connotation as well. Information
includes all of the messages passed back and forth to
trigger all of the computer's operations. In these two
respects, computer information is different from human
information.
|
|
|
Markup
languages like html contain text mixed with instructions on
how to treat that text &endash; instructions and data mixed.
They are not alone in this. Word processor files use markup
languages, and all files contain headers and other
information telling where they begin and end and how to
interpret them. On the computer, there is often no
difference in form between instruction and data.
comment
|
|
|
Within the computer, information often has no occasion, no
point or time of origin or syntactic or semantic context.
When we ask a friend a question, we know where we are, have
an idea of what he or she knows, how the answer might relate
to the larger conversation we are having, etc. This
"postmark" often overshadows the information telling us
whether it is truthful, ironic or a joke. Computer
information often carries no such postmarks, which, as we
know, is partly why it can be dangerous.
|
|
Hypertext
& computers Information processing
|
We can draw some initial contrasts between computer
information and human information processing. For us,
information is most often a relationship of foreground and
background. The background is what we know or presume to be
true, against which something contrasts. Information is
something that relates strategically to a set of
expectations: it creates, confirms or refutes them.
Information is semantic: data that has some significance to
a story. We acquire new information against a background of
experience and we process it by checking it against that
background. Computers do not have such a systematic routine
for accepting new information: computer data does not exist
with reference to stories.
|
|
|
For people, information often involves the matter of
inspection, interpretation and decision making. It involves
shifting the conceptual frame. My friend's car in the
driveway is a sure sign that he has returned from his
vacation. That is a link made by me, a story I am making
using evidentiary material. Thus, information belongs to the
receiver. As a viewer, I inspect the information and I ask
what it means to me. A law becomes information when I ask
myself why it was enacted, and decide whether I will comply
with it. If the data do not seem to cohere, it is I who must
fish around for or create a new concept by which to order
and glue the data.
|
|
|
Information also can and often does operate as message, as
when I read the sign and react without conscious
consideration. At the same time, if your messages can
anticipate my interpretive questions and elicit a response,
they can manipulate or program me. If you are able to
enforce an interpretive frame, we have distorted
communication of the sort Frankfurt school philosopher
Jürgen Habermas addresses. 11
|
|
|
Finally, unlike computer information, manmade text is
produced in a particular place and time and has perceivable
markers to that place and time. Historically, it has been
produced as narrative, a stream of words from speaker to
hearer or writer to reader. If there is information, it has
often been through the pen of the writer and as reliable as
the writer himself. This occasioning of text is
characteristic of pre-computer communications, and it is
being reasserted in forms like email and newsgroups.
|
|
Hypertext
& computers Summary on
interpretation
|
Computers are a particular technical and cultural formation.
We have beliefs about them that condition our use of them.
Within computers, programs determine the interpretive frames
according to preplanned, rule-based procedures, which are
typically predetermined or closed rather than open ended.
Since Hypertext emerged from computer information,
Hypermedia models of communications have been very much
influenced by computer ideology, toward models of
information that are objectified and decontextualized.
|
|
|
Thus, the computer model of information when applied to
humans is tendentious in three respects:
- 1)
information as data versus as a figure&endash;ground
relationship,
2) information as programming, versus as an output of
interpretation,
3) information as objective reality, versus information
as narrative.
- Most
particularly, these models fail to adequately engage the
most important question: intelligibility
|
|
|
In short, tendencies have converged to produce an ironically
naive view of information and interpretation within
hypertext theory. These include the ideology of computers as
scientific: particularly in terms of the apparent clarity
and objectivity of computer data, the belief in the
possibility of universal knowledge, the confusion between
text and work. The view of information they present is
mechanical, not human. Not only have these been the views of
readers; they have also been the views of the creators of
hypermedia communications, and in that sense they are
self-fulfilling.
|
