Or: Why?
At the beginning of the semester I posted the following as my plan
for a special topic area presentation:
For a special topic, something along these lines:
How can machine based tools/information resources be used to augment
human brains (for example, to do lists for remembering things you need
to do but manage to forget) and what are the issues of classification,
representation and accessibility that help or hinder the effective use
of such tools? As I get closer and pick up a few things that broad
area will probably narrow somewhat.
The notion of augmenting human brains quickly led to the writings of
Vannevar Bush and Douglas Engelbart. Ted Nelson followed. Eric
Drexler provides cogent overviews of their issues and presents a clear
justification for hypertext.
Given this tumble towards hypertext I decided that my project should
be a hypertext system that demonstrates some of the benefits and
problems in hypertext systems. I already had warp available from
previous experimenting. I realized, initially, that it would provide a
good system for the required glossary. After some experimentation
I discovered that placing all the content either at the top entries into
warp or inside the warp database would provide an effective
demonstration of several aspects of hypertext:
- Navigational confusion
- Author and authority confusion
- Unexpected discovery of interrelations between things
- The ability to make copious references and the value such "context
generation" provides
Warp also provides a measure of interactivity that can help the
readers make the abstract concepts displayed in the project more
concrete.
[ Top ]
Getting warp to work in the context of the project required some
changes:
- A keyword index of the definitions was created with a search
interface to find the definitions that include the same terms or to find
terms that were somehow associated with the given terms.
- The method by which warp parses text needed to be changed to allow
for external linking. Without external linking it was impossible to
reliably reference external resources from within any text that had
been sent through the warp engine (because it could accidently nest
<a href> tags). The parsing process was adjusted so
segments of the text can be wrapped in tags to indicate only those
segments get sent through the engine.
These changes make warp more usable, but still leave some significant
problems which should be addressed:
- The search system does some probabilistic weighting of words
in the definitions that can sometimes cause a word that is present
in the dataset not to be found. This is a design feature (not a bug)
because it is beneficial for the large datasets for which the indexing
system was designed [mysql_freetext] but can have some weird effects on
smaller datasets (like this one).
- The search index is only of the definitions, not of the words.
This means that if you are looking for a definition for a particular
word you may not be able to find it if the definition does not include
the word. In that situation the glossary word list can be more
helpful.
- The word list interface to the words stored in warp--used by the
glossary--is cumbersome beyond about 15 words or so.
- Although warp keeps older versions of definitions, no facility is
provided to make comparisons between versions. A complete system would
allow side by side comparison, display of differences (like
the diff command in some operating systems) or both.
[ Top ]
The project as a whole has some problems as well:
- The system is perhaps too good at inducing navigational
confusion. Although external references launch new windows, there
really isn't enough screen real estate to follow outside references
while still maintaining a sense of understanding of the text from
which the reference originated. Xanadu and the memex try to address
this issue with multi-page displays.
- There's too much to cover. The scope of this project was made
intentionally broad to indicate the many issues involved and provide a
launching point for further reading. While doing that, however, there
is confusion about the goals: is this is a reference, criticism, a
demonstration or general exposition? On the other hand, unless the big
picture is at least slightly drawn, the impression left is incomplete:
hypertext, especially as we see it in day-to-day life, is fairly
mundane; enhancing knowledge evolution is not.
- Hypertext, by its very design, throws apart any notion of a linear
read of a thesis method, 3.2 paragraph style of writing. It is
difficult, to say the least, to ensure that the reader understands and
receives the intended message.
So, to sum: The goal of this project is to provide a broad overview
of why there is hypertext, who thought about it, what it implies, and
its benefits and problems. At the same time it is hoped that it will
induce something of a gestalt understanding of the gee-whiz factor
that could still be present in discourse related to hypertext and
the augmentation of knowledge processing. If I've fallen short it is
because in the process of my research I have been pulled off my thread
by so many interesting things.
Previous is Howto
This is Rationale
Next is Content
|