Information Flow 2.2 ~~ February 2003 ~~ Information Visualization 2007

Ramana Rao's INFORMATION FLOW ~~~ Issue 2.2 ~~ Feb 2003 ~~~~

~~~ IN THIS ISSUE ~~~ February 2003 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

* Introduction
* Immanuel Kant, but Kubla Khan
* Information Visualization 2007
* Links: InfoViz, Power Law & Weblogs, Aphophenia, XFML

~~~ Introduction ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Here we are at the 10th issue of IFlow (abbreviation is
unavoidable). And I haven't yet specifically talked about
Information Visualization yet. This is clearly a love of mine,
and it seems befitting of Valentines month to sing its song.

I often get interviewed by journalists, analysts, or even former
colleagues in the research community on commercial prospects for
visualization. Some days, I freeze awaiting a definitive
statement to drop down from the heavens into my head (you would
think this would have happened by now). In this issue, I take a
stance. You are welcome to tell me what you think. Feel free to
save the varnish.

~~~ Immanuel Kant, but Kubla Khan ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Concepts without percepts are blind;
Percepts without concepts are empty

-- Immanuel Kant (with a twist)

In Xanadu did Kubla Khan
A stately pleasure-dome decree:
Where Alph, the sacred river, ran
Through caverns measureless to man
Down to a sunless sea.
So twice five miles of fertile ground
With walls and towers were girdled round:
And there were gardens bright with sinuous rills,
Where blossomed many an incense-bearing tree;
And here were forests ancient as the hills,
Enfolding sunny spots of greenery.

-- Samuel Coleridge

Kant was trying to unify empiricism and rationalism, but I twist
the quote to serve my own purposes. Not to spell it out utterly,
the first line in my version is about the need for information
visualization, and the second, graphic design run amok. The
other, Coleridge, is reported to have been on drugs and dreaming.
*See* how words invoke machinery in our heads to create a real
virtual world. Twist that and ask how could visualization *not*
succeed? Feel free to accuse me of being on drugs and dreaming.

~~~ Information Visualization 2007 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I've been pursuing the commercialization of Information
Visualization for over ten years now. And it may well be another
ten years. Clay Shirky, whose articles on a wide range of topics
I quite like, authored the Sept 2002 issue of Release 1.0 on the
topic of Information Visualization. He quoted me in the final
line of the issue:

"Everyone thinks revolutionaries are impatient, but they are
actually extremely patient. We are so stubborn we will keep
at it 'til it happens."

What I might have emphasized is a little less about zeal and
revolutionaries (and me), and more about the reason why patience
is necessary. A real revolution causes some kind of
transformation, and it is very hard to change people and even
harder to change social structures that have set.

Information Visualization, as originally focused on at PARC, was
really seen as the next generation of user interfaces. We saw
particular visualizations as "wide widgets," a new class of user
interface devices that were wide in two ways. The lesser,
already somewhat captured in the widget concept, is about wide
applicability. The more meaning is that these devices widened
the interface between the person and the computer, thus enabling
the person to deal with much more information and use limited
attention more effectively.

Troubles arise not just when commercializing a completely new
paradigm, for example, the graphical user interface, but also for
the next generation. One trouble for Next is that Last is
already entrenched, as Steve Jobs discovered about the niche of
artfully-designed, tightly-integrated PCs.

The basic graphical user interface paradigm was pretty much set
by the time we started to look for commercial paths. It was
owned by Microsoft. Windows 3.1 had shipped, and by the time,
we'd worked through various attempts at Xerox, Windows 95 had
shipped.

It might seem that the answer would be obvious. Work with
Microsoft. And indeed, one of my colleagues at PARC, did go to
Microsoft because he believed that it would be the right path to
get the ideas into the mainstream. But he hasn't had greater
success really.

The problem is not that Microsoft is evil or that they lack the
talent to innovate. Rather, once you've achieved complete
domination at one layer, there is little incentive to innovate at
that layer. Your attention naturally flows to other battles
where your franchise can be parlayed into winning neighboring
estates.

Setting aside that the graphical user interface paradigm was
frozen, another even greater factor stalled UI innovation. We
stopped worrying about bloated desktop applications and the
painful textual, mechanical user interface as it had become,
altogether.

We started spending more and more of our time on the Web and
using communication tools. All of our attention flowed
elsewhere. We were willing to drop back considerably in
interface quality for many years because of the rich sources of
information and knowledge, new services, connections to other
people available through the Internet. Only now are we getting
back to considering simpler and richer ways of interacting with
content, services, people.

Looking forward, there are two paths to mainstreaming information
visualization. One is captured in the ever seductive hope of a
"killer application." This path focuses properly on the fact
that visualization (like other user interface technologies) is
just an ingredient technology, not an application in itself. The
trick is finding an application that is not possible without
visualization or else is significantly improved by it. Of
course, innovation at application level faces equivalent
challenges of getting something new adopted.

I can imagine a new desktop application taking off, for say,
current information awareness, information sensemaking, or
exploratory data analysis. Probably not to happen from Inxight
(for basic business reasons again), but there are mainstream
applications out there now that could wildly succeed. Map of the
Market, perhaps. Groxis, perhaps. Others?

The other path involves a more gradual diffusion. Say like the
10 year diffusion of graphical user interfaces from Xerox Star in
1981 to Windows 3.1 in 1991 in which every two years got you
roughly a power of ten in number of users such that 1000 users
climbed to (say) 100 million user. The diffusion would happen
through multiple applications, some specialized with smaller user
populations, others more general with larger user populations.

Each generation of the idea diffusing creates a population of
users captured by the idea, which becomes the platform upon which
another round of applications are forced or persuaded to
incorporate the ideas. Over time by accumulatation, we arrived
at a day where information visualization shows up everywhere that
it makes sense.

The killer application idea assumes mainstream application draw
so much attention that everybody gets converted rapidly, whereas
the slow diffusion model most naturally works by starting with
highend applications and then being filtered and refined down
into more broadly used applications. I think the first path
could happen still, but if not, the second path will lead to the
spread of information visualization, anyway.

So here is my prediction. In 2007, not quite a dozen years after
the commercial introduction of Inxight Star Tree (original called
the Hyperbolic Tree) and University of Maryland's Treemaps, there
will be 100 million users of information visualizations on a
near-daily basis. And in 2008, $1 billion of revenue will be
attained by applications (including Web Services) that have
visualization as a defining feature.

When that future comes, there will be a few "wide widgets" that
are used in many, many applications. There will be say, twenty
widgetesque techniques that are custom built in many
applications. And there will be some highly specialized
visualizations used in particular applications. Okay, I can't
avoid saying it will be a power law distribution.

If you think this sound outlandish, think about the current
distribution of communicational graphics techniques today. Since
the invention of the Cartesian plane, we've seen many new visual
representations. Some like the scatter plot, pie charts, bar
charts, so on, you see everywhere and can read with ease. Others
become part of the common visual language of particular
communities. For example, scientist, economists, engineers all
specialized versions of some of the generic types. Finally there
are some visual communications designed and used only one time.
Like the famous diagram of Napoleon's march into Russia.

So too will it be with interactive visualizations. It cannot
*not* happen.

~~~ Links ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

~> See and Go Manifesto
~> http://www.ramanarao.com/papers/seeandgomanifesto-acm.pdf

This "reflection" was written in 1999, twelve years after Stu
Card started information visualization research at PARC. It
was a "manifesto" calling for the next generation. Also it
summarizes the key principles of wide widgets.

~> Information Visualization: Graphical Tools for Thinking about Data
~> http://www.edventure.com/release1/abstracts.cfm?Counter=5866492

The Sept 2002, Release 1.0 issue written by Clay Shirky. You
can read the abstract, but the full issue costs $80. Oh well,
maybe I can convince Esther or Clay to share this article.

~> Power Laws, Weblogs, and Inequality
~> http://shirky.com/writings/powerlaw_weblog.html

A good article, also by Clay Shirky. Besides providing
background on power laws, this article has explains what kind
of lock

~> Review of Gibson's new book, "Pattern Recognition"
~> http://www.salon.com/tech/books/2003/02/13/gibson/index.html?x

I haven't read Gibson's book yet, but this review is
interesting. Another word to add to the coincidence list.
Apophenia is "the spontaneous perception of connections and
meaningfulness in unrelated things." As the revier says,
"Recognizing patterns that aren't actually there."

~> XFML, eXchangeable Faceted Metadata Language
~> http://xfml.org/

Facets (see IFlow Aug 2002) meet XML. I've known about this
new standards effort for a while, but haven't had the time to
really understand it. Maybe one of you can make some comments
on this that you would allow me to include in a future issue.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Ramana Rao is Founder and CTO of Inxight Software, Inc.
Copyright (c) 2003 Ramana Rao. All Rights Reserved.
You may forward this issue in its entirety.

See: http://www.ramanarao.com
Send: [email protected]
Archive: http://www.ramanarao.com/informationflow/archive/
Subscribe: mailto:[email protected]
Unsubscribe: mailto:[email protected]