From: Gregory B. Newby <gbnewby@alexia.lis.uiuc.edu>
Subject: Re:  Sources of the concept of "information as space"?
Date: Wed, 11 Dec 91 11:30:04 -0600



Thom Gillepsie <thom@dewey.soe.berkeley.edu> writes:

>I know that Gibson coined the term CyberSpace, and I know that Vinge first
>described an 'information space' in True Names, but I wonder if there are oth
>earlier instances of information as space in the 'popular' press. I'd be
>particularly interested in  'information as space' concepts in other cultures.
>Email me with suggestions and ideas and I'll post if there is interest. Thank

Hi, Thom.  I think we corresponded a few months ago, and I never
followed up.  My apologies.

I've done a whole lotta thinking about information space, mostly
in the context of information science (in general) and information
retrieval systems (specifically).  I've seen pretty well Zip
in the popular press.

However, Information Space is a phrase that exists in several academic
literatures (psychology, information science, and business at least),
but has not been defined anywhere that I've seen -- even informally!
(Thanks for the Vinge tip: I'd better read that book!)

Below is the conceptual framework I'm using to think about information
space.  I'll try not to get long here, as this topic is a basis for my
dissertation work, and I'm also trying to write an article or two for
it.  I'll work backwards:

information space is a particular incarnation of cognitive space,
	usually associated with databases or information retrieval
	(IR) systems.  In this type of space, the human agent
	navigates through the information space, and the 
	space is relatively inflexible.  That is, things tend
	to stay put in an information space (as opposed to cognitive
	space, below).

cognitive space is the necessary medium in which cognitive movement
	(see below) takes place.  It consists of two things:
	(1) concepts and (2) relations among the concepts.
	This is basically the framework in which "knowledge" can
	be said to exist.  The power of this scheme is that in
	a space, things can all have some relation to each other --
	unlike frames, scripts, hierarchies, and most other
	representation schemes in which many things have no
	measurable relationship.  The problem, of course, is
	that it's pretty vague.  (However, I have implemented
	a sort of cognitive space for IR, so there's hope ;-)

cognitive movement is a fundamental human condition.  It's what
	happens when our understanding of concepts or relations
	among them change.  Cog. movement might be learning,  or
	it might be forgetting.  Cog. movement can result in 
	new knowledge, changes in opinion, identification of
	gaps in knowledge, etc...

How does cognitive movement come about?  By communication,
	of course!  Communication is the transmission of messages
	between two or more entities.  Very briefly:  

		In the case of human communication
	in real time, there is the ability for all parties'
	cognitive space's to interract:  cognitive change can
	take place for all people.  Of course, what change takes
	place is largely dependent on the cog. space that 
	existed before the communication.  Hence, there is a
	"negotiation" between the cognitive spaces of the 
	parties involved.
		
		In the case of IR systems (accessing a database),
	and most of virtual worlds, and reading a book, and
	other situations when "communication" takes place between
	an active human and some sort of pre-existing source, 
	only the human's cog. space can change.  Hence, the 
	human "navigates" through the information space of the 
	system.  In VR, there's potential for the information space
	to change, to adapt to a particular user.  Hence, there's  
	potential for doing more than just navigating through
	an existing space.

In a small nutshell, those are some of my thoughts on information
space.  I ended up building a spatial information retrieval
system, in which you navigate among keywords and documents from
a database (things are related based on similarity:  things which
are more similar (as measured statistically) are closer together).
It works OK, but I need more relations among concepts and some
additional navigation cues.  I'm in the midst of data analysis
collected for my evaluation.  This is implemented as a 3D
information space, but viewed on a regular Personal Iris.  Navigation
is via mouse or PowerGlove.  More details forthcoming...

Perhaps this work will be an early incarnation of "cyberspace."

Perhaps it will revolutionize information retrieval.

Perhaps not....

Please don't flame me too horribly on this stuff.  It's
just a proposal, and hasn't been largely validated or tested.  However,
I did draw on several existing literatures.  Your comments and
criticisms are most welcome!

-- Greg Newby
   Grad. Sch. of Lib. and Info. Sci. / Nat'l Cntr. for SuperComputing App.
   U. of Illinois at Urbana-Champaign 
gbnewby@alexia.lis.uiuc.edu	"Curiouser and curiouser"  -Alice