Designing The Future of Information: The Internet Beyond The Web

DESIGNING THE FUTURE OF INFORMATION
The Internet Beyond the Web
Current computer science operates with models of information,

networking, and databasing that were conceived in the mainframe
era and cannot serve the needs of a truly connected world.
Two unique initiatives—the “Information Commons” of

MAYA Design and “Internet Zero” from MIT’s Center for
Bits and Atoms—offer simple, proven foundations for the
intense complexity of a global information economy.
a harbor white paper

September 2005
Harbor Research, Inc.

BOSTON | SAN FRANCISCO
Designing the Future of Information

Overview
“The Information Commons” (Maya Design, INC.)
4 Global information space for everyone. Overcomes the Web’s “information
islands” to offer scalable interoperability for the pervasive era. Enables civic
computing with public data but protects intellectual property for commercial use.
4 New information technology, next-generation capabilities. Simple, scalable
architectures for information devices, data storage, and information itself. Enables
sophisticated fusion and visualization of data that will never be possible with the
Web alone. The result of $50M in R&D to date.
4 Global P2P system. Fully distributed computation and storage/retrieval.
4 Proven in commercial, government, and non-profit implementations.
“Command Post of the Future” helps run U.S. military operations in Iraq today.
4 Free access and browsing/authoring app. Java-based “Civium Workbench” in
public beta by the end of 2005. Powerful geospatial features, among many others.
4 Works with the Web and legacy systems. Information can be published into
and out of the Commons from the Web and other systems that use the relational
databasing and client-server model that the Commons replaces.
4 Could replace the Web. Future-proofed design principles and “emergent
intelligence” potential could make it the Web’s replacement for the 21st century.
4 Multi-disciplinary team and approach. Research lab/design consultancy spun out
of Carnegie Mellon University. Includes luminaries from the fields of engineering,
information visualization, interface design, and human sciences.
“Internet Zero” (MIT’s Center for Bits and Atoms)

4 Emerging standard for connecting anything. Replaces today’s many so-called
connectivity standards with a truly interoperable “Morse Code for the Internet.”
4 Embeds IP in the physical world. Back-to-the-future concept extends the
scalable, agnostic protocols of the Internet to any device, even non-electronic ones.
Incorporates Internet Protocol into physical infrastructure.
4 Multi-disciplinary team and approach. Faculty from many MIT departments.
Includes one of the fathers of the Internet itself.
© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/
info@harborresearch.com • 800.595.9368 • 415.615.9400

W
hen it comes to preparing for the global

information economy of the 21st century, most
people assume that “the technologists are taking
care of it.” They take it on faith that the best
possible designs for the future of information will
emerge from large corporations and centralized authorities. But those
are big, unfounded assumptions. In fact, most entrenched entities are
showing little appetite for radical departures from current practice.
Yet current practice will not serve the needs of a genuinely connected
world. This paper examines two real-world initiatives based upon
new thinking about information management and networking. They
demonstrate that it is possible to migrate gracefully to scalable
architectures designed for the era of pervasive computing.
Introduction
The Pervasive Internet: IT means what it says
For quite a few years now, Harbor Research has focused most of its research and consulting
on what we call “the Pervasive Internet”—the convergence of pervasive or embedded com-
puting with the packet-switching “network of networks” called the Internet.
We prefer “the Pervasive Internet” over other terms in common use—notably “M2M,”
which usually stands for “machine-to-machine”—because it captures the profound enor-
mity of the phenomenon: the world on the Internet, the Internet in the world.
Pervasive computing—also commonly called “ubiquitous” or “invisible” computing—usu-

ally refers to digital microprocessors and sensors embedded in everyday objects. But even
this makes too many assumptions about what the pervasive phenomenon will be. Encoded
information in physical objects is also pervasive computing—even without intrinsic com-
puting ability, or, for that matter, without being electronic at all. Seen in this way, a printed
bar code, a CD or DVD disc, a house key, or even the pages of a book can have the status of
an “information device” on a network.
Lucas, “The Trillion-Node Network,” p. 3.


But very few people are thinking about pervasive computing on that level. We all casually
repeat phrases like “the network is the computer” without really considering the implica-
tions. “Pervasive computing” should automatically be understood as “networked pervasive
information and computation,” but it isn’t. The nature and behavior of a truly distributed
global information system are concerns that have yet to take center stage—not only in busi-
ness communities, but in most technology communities, too.
The “Information Commons” and “Internet zero”

In the following pages, we examine two unique and important initiatives from people who
are thinking on the scale that the Pervasive Internet deserves—the “Information Com-
mons” of Pittsburgh-based MAYA Design, Inc. (http://www.maya.com/), and “Internet 0”
(I0) from MIT’s Center for Bits and Atoms (CBA) (http://cba.mit.edu/).
In our years of work on the pervasive phenomenon and its real-world effects on business, we
have not encountered more compelling visions of the future of information than these. The
creators of both these initiatives have taken great pains to keep their thinking as simple, gen-
eralized, extensible, and irreducible as possible. That, as much as anything else, is the theme
of this paper.
The Information Commons and Internet 0 are disruptive technologies. They are not in-
cremental improvements, patches, Band-Aids, or new flavors of what we already do. They
represent a true paradigm-shift in IT thinking. The Information Commons, for example,
looks toward the creation of a universal database to which anyone can contribute, and which
liberates information by abandoning relational databasing and the client-server computing
model. Internet 0 envisions one simple standard, based upon Internet Protocol itself, for
networking any information device (even non-electronic ones), rather than the multiplicity
of so-called standards we have today.
At the same time, taking these initiatives seriously does not mean junking all current IT
practice in one fell swoop. The pillars of present-day information technology will not crum-
ble overnight, nor has the great existing investment in them suddenly lost all value. There
are reasonable, fiscally sane paths for migrating to the future. But migrate we must. The as-
sumptions and practices of the mainframe and PC eras are now decades old and not suitable
for the pervasive computing era.

The “Information Commons” of MAYA Design

Can the Pervasive Internet Be “Designed”?
“The Pervasive Internet” really means the future of information, and that means the future
of civilization. It will require a remarkably agile global network that could comfortably scale
to trillions of nodes—some of them hardware, some software, some purely data, many of
them coming into and out of existence or changing location constantly. Obviously, such a
network cannot be “designed” in any ordinary sense. Certainly, it cannot be designed “top-
down.”
And yet the Pervasive Internet must be designed in some sense. Such a network will easily
be the biggest technical achievement in the history of humanity. Its closest predecessor is
the global financial economy—with which, in fact, it will share vital characteristics. Some
basic design principles must be put in place to guide the
We are talking here about the design growth of a vast, distributed technological organism that
of systems that will evolve and must remain organized as it evolves according to a logic all
its own. It demands that we design not only devices and
unfold according to a logic all their
networks but also information itself in ways not addressed
own—a possibility that will require by current IT.
an expansion of our understanding of
design work. “Most Advanced Yet Acceptable”
For more than 15 years, this has been the underlying pre-
“Information Architecture and the Emergent Properties of
Cyberspace” —MAYA Design occupation of most work done by MAYA Design, Inc.,
a multi-disciplinary technology research lab and design
consultancy founded by three Carnegie Mellon professors
with diverse backgrounds—human sciences, visual design, and engineering. MAYA’s name
is an acronym for “Most Advanced Yet Acceptable”—a phrase borrowed from Raymond
Loewy, who is usually called the father of industrial design. (Loewy was the creator of the
Studebaker Avanti automobile and the Shell Oil logo, among other classic designs.)
MAYA’s co-founder and intellectual leader, Peter Lucas, and his longtime collaborator and
CTO, Jeff Senn, have tried to connect all their work to the overarching agenda of liberating
information from arbitrary restrictions and restraints.
Along the way, they have had many illustrious colleagues and collaborators. MAYA co-
founder Jim Morris spent eight years at Xerox PARC and worked on the legendary Alto
Lucas, “The Trillion-Node Network,” p. 2-6.


computer that inspired the entire era of mouse-driven, windows-based computer interfaces.
He later left MAYA to become Dean of the Carnegie Mellon School of Computer Science.
The late Steve Roth, co-founder of MAYA Design’s sister company MAYA Viz (recently ac-
quired by General Dynamics and renamed simply Viz), was a pioneer in data visualization
with his seminal work at Carnegie Mellon’s SAGE group. Many of Roth’s ideas are funda-
mental to all MAYA software. Bruce Horn, co-designer of the original Finder for the Apple
Macintosh, was another early “MAYAn,” as the company likes to call its people.
Since their inception, MAYA Design and MAYA Viz have won more than $50M in gov-
ernment and corporate contracts to conduct research and create solutions for next-genera-
tion computing and information technology. Their architectural and visualization break-
throughs can be seen in “Executive Command Center,” a commercial product for business,
and “Command Post of the Future” (CPoF), a new system for real-time situational aware-
ness and collaborative decision-making now being used by the U.S. military in Iraq. Retired
Army General Paul F. Gorman called CPoF the “greatest advance in technology for Com-
mand and Control in the past 30 years.”
Much of this innovation has now found its way into MAYA’s “Information Commons,” a
global information space designed for everyone and every conceivable kind of data, with the
scalability and information-freedom needed by the pervasive-computing era.
The Information Commons has been in the lab a long time. It’s now coming out into the
world—by bringing the world into itself. The Commons already contains, for example, in-
formation on all worldwide geopolitical entities, physical features of the Earth, human-cre-
ated structures such as buildings and transportation infrastructure, U.S. Census data, and
much more.
By virtue of MAYA’s unique information-architecture concepts, these datasets can be

mapped much more powerfully than in most commercial GIS applications. And the Com-
mons allows anyone to become a contributor by creating new units of information or by
commenting upon existing ones, all the while clearly separating fact from opinion.
Why Do We Need the Information Commons

When We Already Have the Web?
Before delving into the new thinking that makes all this possible, let’s talk about why it’s
necessary at all. Don’t we already have a vast public information space called the World


Wide Web? Didn’t the Web completely revolutionize human communication? And isn’t the
Web working and scaling quite handsomely?
Almost everyone will answer with a resounding “Yes!” But consider this analogy from
Buckminster Fuller: Suppose you are traveling on an ocean liner that suddenly begins to
sink. If you rip the lid off the grand piano in the ballroom, throw it overboard, and jump
on it, the floating piano lid may well save your life. But if, under normal circumstances, you
set about to design the best possible life preserver, are you going to come up with the lid of a
grand piano?
The World Wide Web is like that piano lid. In a period of great change and tumult, it
worked—in the sense that it kept us afloat. But that does not make it the best possible de-
sign, or qualify it to be something that we should plan to live on forever.
Yet, in the course of one mere decade, the world has become so dependent upon the Web
that most people, inside IT and out, cannot bring themselves to think about it with any
critical detachment. Even in sophisticated discussions, the Web and its key enabling tech-
nologies are usually viewed as utterly inevitable and unquestionable.
IT professionals, for example, rarely talk these days about the need for ever-evolving infor-
mation services that can be made available anywhere, anytime, for any kind of information.
Instead, they talk about “Web services.” Even high-tech business people use the terms “the
Web” and “the Internet” interchangeably without giving it a thought.
But the Web is not the Internet. The Internet itself is a simple, elegant, extensible, scalable,
technology-neutral networking system that will do exactly what it was designed to do for
the indefinite future. The same cannot be said of the Web, which is essentially an applica-
tion running on top of the Internet. It is hardly the only possible Internet application, nor is
it the most profound one conceivable.
Now, after a decade of rampant, unruly Web proliferation, we see that the Internet’s inher-
ent scalability has been both a blessing and a curse. The Internet was able to “give the Web
all the rope it wanted,” and today the Web finds itself trying to be something that it was
never designed to be: the fundamental platform for the future of humanity’s information.
This is the dilemma that MAYA’s Information Commons proposes to resolve by providing
something that really was designed to be such a platform—with the extremely well-thought-
out device- and information-architectures required.
“The Internet of Things,” p. 76.


The Web Consists of information islands
The Web’s Achilles heel does not originate in its browsing software, or markup languages, or
the other superficial aspects that most users touch directly. Those inventions are not neces-
sarily ideal, but they are useful enough today, and they can be replaced over time with better
alternatives. Rather, the Web’s weakness lies with its basic enabling technologies—in par-
ticular, relational databasing and the client-server model—and
the restrictions they place upon structuring, storing, and re-
trieving data.
Demand for interoperability is
growing, and as designers of
The Web stores information in one of two basic ways: utterly
unstructured, or far too rigidly structured. The unstructured information devices work to provide
way gives us typical static Web pages, blog postings, etc., in it, they will be laying the foundation
which the basic unit of information is large, free-form, and for an information system far
lacking any fundamental identity.
vaster than the existing World Wide
The overly structured way involves the use of relational data-
base tables that impose rigid, pre-ordained schemas on stored
Web. Cyberspace will spill from
information. These schemas, designed by database administra- our desktops and extend tendrils
tors in advance, are not at all agile or easily extensible. Mak- into our kitchens, our cars, and our
ing even trivial changes to these schemas is a cumbersome,
expensive process that affects all the data inside them. Just as
clothing. It will be many orders of
importantly, they make deep, inflexible assumptions about the magnitude more complex than any
meaning and context of the data they store. artifact human beings have ever
Both of these approaches to data-structure enforce severe limi- encountered.
tations on the two things you want most in a global, pervasive- “Information Architecture and the Emergent
era information system: scalability and interoperability. Properties of Cyberspace” —MAYA Design
The client-server model underlying the Web greatly com-
pounds the problem. Regardless of data-structure, when you put information on the Web
you put it in a specific physical location—i.e., you upload your HTML or middleware files,
or your database tables, to a specific directory on a specific server. Thus, a Web URL does
not point to information per se; it points to the information’s physical location, which is a
very different thing.
This means that the Web is machine-centric, not information-centric. Information on the
Web is not free to move, and because its life is tied to the life of a physical machine, informa-
tion on the Web can easily become extinct. If a particular server goes offline, temporarily or


forever, all its URLs are dead and the information is unavailable. You can “mirror” your data
so that copies reside on multiple servers, but this does not change the fact that Web-based
information is always tethered to specific locations on specific machines.
All of this adds up to a huge collection of information-islands called the World Wide Web.
Assuming the islands remain in existence reliably, they are still fundamentally incapable of
truly interoperating with other information-islands. We can create bridges between them,
but islands they remain. That’s what they were designed to be.
The latest bridges are called “Web services,” by which Web islands automatically “feed”
information to each other. In terms of convenience for human beings, Web services are a
distinct advance over manually navigating Web pages. You can now set up a browser page
or a desktop applet to “consume” (receive and display) a Web-feed of news headlines, stock
quotes, weather information, package tracking, and so on. And that’s a good thing—a pleas-
ant, time-saving, incremental improvement on yesterday’s way of doing things.
But many people talk about Web services as “the next big thing,” and if that’s true we can
only say: We were hoping for something bigger. Consuming a Web service is exactly the
same thing as loading a Web page. The networked Dashboard Widgets in Apple’s Mac OS
X, for example, are really little mini-browsers set up to hit Web URLs; you just don’t see
them doing it.
Web services are still prey to the client-server rigidity and vulnerability we’ve described,
making them particularly unsuitable for vital public information services. And in order
to concatenate information from diverse sources, Web services depend upon data-tagging
conventions that require yet more layers of prior agreement, schemas, markup, and manual
human administration.
With each additional layer of such engineering and administration, the Web comes closer
and closer to resembling a fantastically jury-rigged Rube Goldberg contraption. The reason
is simple. The Web, with or without “services,” was not designed for a world driven by per-
vasive information flow.
Why Do We Need the Information Commons When We Have Google?

In its earliest days, the Web did not pretend to be the ultimate global information platform
for all time. Its raison-d’être was more idiosyncratic and less glorious than that, though it
was clearly a remarkable innovation—Everyman’s first hands-on experience of a global data
network. Back then, hard-coded links were what added meta-value to the unstructured in-


formation stored on the system. Human beings indexed the Web manually, with their eyes
and hands, and well-maintained “jump pages” were the most valuable Web destinations.
Those days didn’t last long, of course, but it’s important to remember what the Web was
originally designed to be. Soon, automated indexing via search engines became essential.
Those engines had names like AltaVista and Lycos, and when you searched for something,
you got about 50,000 badly prioritized results, and you spent half your day barking up the
wrong tree.
Then came Google. It was started by some very smart people, and it produced vastly bet-
ter results than the competition. Today, “Google” has become a verb and the company is
among the most bankable in the world. If Google, or something else like it, had not arrived,
humanity might have felt significantly less confident about depositing virtually all of its pre-
cious information on the system called the World Wide Web.
But though it is a remarkable achievement, Google remains a blunt instrument, not a pre-
cision tool, for dealing with the world’s information. No matter how clever its indexing
schemes and searching algorithms, Google cannot get around the fact that the Web does
not store information in a fashion conducive to sophisticated retrieval and display. The Web
can’t genuinely fuse data from multiple sources, and so Google can’t build an information-
model that visualizes the answer to a complex, multi-dimensional question.
Data Fusion: Information Wants to Be Free

Information on the Web is not free (and that’s free as in “freedom,” not free as in “free
of charge”). In fact, it’s closer to being imprisoned than it is to being genuinely liberated.
Thanks to the client-server model, Web information is not free to move anywhere at will.
And thanks to our present information architectures, it’s not free to merge with other in-
formation.
What would truly liberated information be like? It might help to think of the atoms and
molecules of the physical world. They have distinct identities, of course, but they are also
capable of bonding with other atoms and molecules to create entirely different kinds of mat-
ter. Often this bonding requires special circumstances, such as extreme heat or pressure, but
not always.
In the world of information, such bonding is called “data fusion.” The special circumstances
that provoke fusion might be nothing more than a user asking a certain multi-dimensional

10
question. And the newly created matter would be an easily perceivable, manipulable, or
mappable “model” of the answer to that question.
Here’s a geospatial example that already exists. In Pittsburgh, PA, USA, community devel-
opment groups (in partnership with Carnegie Mellon University and the Brookings In-
stitution), are importing a great variety of city-related datasets into MAYA’s Information
Commons—from property assessment and land-use information to statistics on building
permits and crime. Considered in isolation, none of these data would yield much more than
we might get with conventional Web deployment. But MAYA’s leaders have a longstanding
interest in civic computing and enhanced access to public information, so the Commons
also already contains the location of every toxic waste site in the United States, as tracked
by the EPA.
Remember that all of these datasets were developed independently by entirely different
agencies, and never consciously intended to be used together. But they all contain location
information, often in the form of latitude and longitude. In the Information Commons,
these shared attributes become fusion-points by which one can model the geographic rela-
tionship of, say, building permits to toxic waste sites, or land-use to crime. Any other shared
attributes, such as calendar dates, could also be fusion-points between these datasets. And
any other dataset of interest—the incidence of certain types of cancer, building-code viola-
tions, recreation areas, etc.—could also be fused into a visualizable information-model of
the city or one of its neighborhoods.
Facilitating data fusion is one of the fundamental purposes of MAYA’s Information Com-
mons. Most people have had little, if any, experience with it, and so its power and impor-
tance are often difficult to grasp. The ability to perform true data fusion is the holy grail of
information systems because it allows not only patterns but a whole higher order of intel-
ligence to emerge from large collections of ordinary data. The implications for research in
almost any area of human inquiry are obviously immense.
The Commons uses an entirely new, schema-free database technology (we’ll discuss it in a
moment) that avoids the confinements and limitations of relational databases. It allows data
to maintain their fundamental identity while bonding freely with other data. Furthermore,
the attributes of any data object can be changed or extended without changing any other.
Even so, datasets rich in potential fusion-points are most desirable. Thus, there are best prac-
tices and recommended minimum attributes for most Commons data. In the case of build-
ing permits, for example, MAYA is working with the Brookings Institution to assess the

11
practices of many major cities in order to produce an ideal building-permit specification that
reflects the accumulated wisdom of the profession at large. Cities and towns that adopt this
specification will know, at a minimum, that their Commons-stored data will fuse well with
building-permit data from anywhere else.
Data in the Information Commons have strong inherent identity and a format that oc-
cupies a middle ground between unstructured and overly structured. A recommended set
of attributes for a Commons data object or whole dataset is not the same thing as the rigid
schemas of relational database tables, and does not lead to walled-off, non-interoperable in-
formation silos.
Isn’t the information commons the same as the Semantic Web?

Recognizing the need to share and fuse data from many diverse sources, the World Wide
Web Consortium (W3C) launched an initiative some years ago called “the Semantic
Web” whereby information stored in static pages and relational databases on servers can
be “Webbed”—that is, made newly fluid and portrayable on the Web for both humans and
machines.
Many people who hear about MAYA’s Information Commons assume that it is another ver-
sion of the Semantic Web, and thus redundant and unnecessary. The two visions may seem
to have the same goal, but in fact they are profoundly different.
The Semantic Web proposes to liberate data not by re-thinking basic information archi-
tecture or the machine-centric networking model, but by introducing a complex new layer
of metadata (information about information). This is done with a markup language called
RDF (Resource Description Framework) that describes the potential meanings and rela-
tionships (fusion-points) of a given piece of information in various contexts. RDF is the
Semantic Web’s way of trying to add smartness to the inherently dumb Web.
The Semantic Web does recognize the dilemma presented by the Web’s limitations, and it
has many passionate proponents—perhaps precisely because it does not propose any par-
ticularly radical or disruptive steps. However, the Semantic Web has had very few real-world
deployments. Besides being forbiddingly arcane and captive to existing Web technologies,
RDF metadata often grows significantly larger than the data it is trying to describe.
That last point has important negative implications for pervasive computing. Tim Bern-
ers-Lee, the inventor of the Web, has written that “one of the main driving forces for the
Semantic Web has always been the expression, on the Web, of the vast amount of relational

12
database information in a way that can be processed by machines.” Even though he wrote
that in 1998, Berners-Lee surely understood that “machines” could not mean simply hu-
man-operated PCs with the horsepower and the gigabytes of memory and storage required
by billowing clouds of metadata around every single information object.
Berners-Lee went on to say that “the goal of the semantic web is to express real life”—a com-
pletely admirable sentiment, but the “real life” of a global information space will involve the
networking of many billions of comparably modest information devices, right down to such
small but vital things as smoke detectors and burglar alarms.
At the end of the day, the Semantic Web takes a step in the right direction, but it takes that
step in a cumbersome way, without re-examining the Web’s underlying assumptions. It tries
to fix a broken system with more engineering. The Information Commons, by contrast,
imagines an Internet-connected world beyond the Web, and it does so with architectures
that embrace every conceivable information device, however large or small.
Like the Internet itself, a future-proofed global information system should be based upon
fundamentally simple, extensible ideas. The Information Commons represents a shift to
that sort of new paradigm. The Semantic Web represents an attempt to salvage the old para-
digm.
“Infotrons” and Information device Architecture (IDA)

Radical new thinking about information technology must begin at the most basic levels,
with new conceptions of information devices and of the fundamental unit of information
itself. MAYA’s leaders attempt to future-proof their innovations in these areas by making
the fewest possible assumptions about the nature of networked objects and the data they
carry or process.
Thus, the company takes a much broader, all-encompassing view of information devices (or
“infotrons,” as they’re called) than is typically found in conventional computer science. In-
fotrons are the basic building blocks of MAYA’s Information Device Architecture (IDA).
All infotrons have inputs and outputs, but little else about them is taken for granted. The
essence of an infotron is completely abstracted from its real-world embodiment. In IDA,
infotrons are fungible—that is, mutually interchangeable. From outside a properly designed
infotron, there is no way to know whether it is hardware or software, because it doesn’t
matter. Infotrons may, and often will, be capable of computing, but they don’t have to be.
Berners-Lee, “Web Design Issues: What a Semantic Web Can Represent.”

13
Indeed, an infotron doesn’t have to be electronic at all. If it’s capable of having information
encoded within it, in any form, it’s an infotron.
Infotrons also transmit information in a significantly different way from most distributed
computing systems popular in industry today. IDA was designed from the ground up to
let individual components communicate via “asynchronous message-passing,” which means
that IDA-based designs tend to get better, not worse, as they scale. IDA encourages and re-
wards parallel computation, which is almost the same thing as “pervasive computing.”
A Representation of “Infotrons” and MAYA’s Information Device Architecture (IDA)
In MAYA’s IDA, virtually any object that contains information—whether hardware or software—is an
“infotron” that can be connected to any other.
Source: MAYA Design, Inc.
In the past, the art of parallel computing has involved solving problems fast by getting lots
of machines to work on what is essentially the same computation. With IDA, the art is
to get machines to work together productively when their users have several different, in-
dependent, but overlapping information needs. People can work together in the Informa-
tion Commons, and so parallel computing in the Commons is less about crunching one big
problem than it is about facilitating complex collaboration.
The U-form: A Universal Data Container

The bit, the byte, and later the packet made possible the entire enterprise of digital com-
puting and global networking. Until the world agreed upon these basic concepts, it was

14
not possible to move forward. The next great step in IT—completely fluid information and
fully interoperating devices—requires an equally simple, flexible, and universal abstraction
that will make information itself truly portable in both physical and information space, and
among any conceivable information devices.
A Representation of the Structure of U-Forms in MAYA’s VIA Repository
The attribute-value pairs of a u-form are arbitrarily extensible by users (though there are best practices
that will maximize useful data-fusion). Note that a u-form can reference any other u-form by using its
UUID as the value of an attribute.
As we have seen, chunks of information that lack structure or have too rigid a structure
tend to live in isolation, unable to discover and bond with others to create new information.
What’s required is a data-structure that provides unmistakable identity for every chunk of
information but also facilitates flow and fusion. To this end, MAYA has built a new infor-
mation architecture upon a universal data container called the “u-form.” All data and data-
relationships in the Information Commons are stored in u-forms, not in relational database
tables.
Lucas and Senn, “Toward the Universal Database,” p. 1.

15
The u-form is MAYA’s extension of the “e-form” proposed by the late Michael Dertouzos,
longtime director of MIT’s computer science labs, who recognized that computing systems
need a simple, extensible, and universal way to represent data. An e-form is simply a list of at-
tribute-value pairs. Such pairs are not uncommon in computing, but until MAYA’s u-form
they have not been used as the fundamental unit of information in a system.
To Dertouzos’s e-form MAYA has made one very important addition: a universally unique
identifier (UUID). Thus the “u” of u-form. The UUID travels
with the u-form’s data wherever in the world it may go, and in [We] take the radical, but ultimately
whatever device it might find itself.
necessary, position that there is
Universally unique identifiers (UUIDs) only one space of interoperating
Why does the new universal data container need a UUID? We applications; therefore, at the
all recognize the need for unique identifiers in many aspects of
our lives and work. ISBN numbers for books, Social Security
end of the day there must be only
numbers for U.S. citizens, license plate numbers for automo- one identity space. Until we come
biles, and so forth, allow us to refer unambiguously to a single to terms with the implications of
object in some large population of objects.
this simple claim, we will remain
The aforementioned identifiers, however, are unique only in
their own little worlds, not in the world at large. They are not
doomed to islands of context and
universally unique. They operate within islands of reference, systems that cannot interoperate.
and they do not interoperate with other “islands.” They allow “Mobile Devices and Mobile Data—Issues of Identity
us to point unambiguously at an object in a given “namespace,” and Reference” —MAYA Design
but nowhere else.
In an unconnected or minimally connected world (a world of information islands), this

works well enough most of the time. But MAYA is designing a system for a genuinely con-
nected world in which there are no artificial barriers between pieces of information (data
objects). In such a world, there is only one identity space—namely, Cyberspace—and thus
the entities (information objects) that inhabit that space must be universally unique, not
simply locally unique. Thus the absolute need for every data object to have a UUID.
Establishing UUIDs for every piece of information in the world, however tiny or granular,
will sound impossible to most people. In fact, technically speaking, it’s comparatively easy
Dertouzos, The Unfinished Revolution, p. 50.
Lucas and Senn, “Toward the Universal Database: U-forms and the VIA Repository,” p. 4.
Lucas, “The Trillion-Node Network,” p. 6.

16
to generate identifiers that virtually eliminate the possibility of “identity collision” for all
time.
MAYA Design did not invent the idea of UUIDs or the various techniques that exist today
for generating them. But MAYA is unique in using UUIDs for u-forms, and thus placing
them at the very heart of radically new information, storage, and device architectures.
The Visage Information Architecture (VIA)

Whereas infotrons are the basic unit of MAYA’s device architecture called IDA, the u-form
is the basic unit of a new approach to information itself, called the Visage Information Ar-
chitecture (VIA). The name comes from MAYA software called Visage, which was an early
manifestation of these ideas. Descendents of Visage include the commercial software tools
we mentioned earlier, such as “Command Post of the Future,” now developed and support-
ed by Viz (formerly MAYA Design’s sister company MAYA
If we can achieve a universal identity Viz before its acquisition by General Dynamics C4 Systems).
the name “Visage” suggests, information visualization
space, we will have set the foundation As has always been one of MAYA’s key concerns.
for universal interoperability. It is
VIA is profoundly different from the information architec-
difficult to see how we could make tures in common use today, which one can see manifested
much progress toward this goal in the schemas of relational database tables. Those inflex-
ible, pre-ordained schemas impose more than rigid structure
otherwise.
upon information. They impose rigid meaning. VIA, like
“Mobile Devices and Mobile Data—Issues of Identity and
IDA, is completely abstracted not only from the devices that
Reference” —MAYA Design
may transport or process data, but also from the schemas
and layers of semantics that confer specific meaning upon
data in specific contexts. VIA recognizes that database designers do not have—nor should
they have—any idea of how specific data might be used in the near or distant future.
Instead, VIA deals with the meaning of data when the data are retrieved for a specific pur-
pose, not when they are stored. In other words, VIA does not make interpretive assump-
tions part of the basic unit of information itself.
Doing this allows any datum to move throughout the system “atomically” as an irreducible
entity, with no pre-defined role or absolute meaning imposed on it from birth. The result-
ing “data liquidity” allows these atoms of information to bond with others to form new
information molecules that could not have been anticipated by the creators of the atoms

17
themselves. In VIA, irreducible units of information (u-forms) are free to play any number
of “roles” and thus to be seen—literally “visualized”—in a wide variety of ways, something
that MAYA calls “polymorphism.”
The VIA Repository: One Universal database

The inexorable movement toward pervasive, distributed computing (“the network is the
computer”) implies the existence of a equally pervasive, distributed store of knowledge that
finally does away with all theoretical barriers between the world’s units of information. The
u-form-based VIA makes this possible with a new kind of database technology in which u-
forms are the “records.” MAYA calls this storage architecture
the VIA Repository. Above all else, the agenda implied
Everything in the VIA Repository is represented by a u-form. by the Universal Database requires
Whole collections of u-forms (which might contain, for ex- a technique for unambiguously
ample, the result of a complex data-fusion) are themselves con-
tained in a u-form, as are simpler u-form relationships. Each preserving the identity of each and
of the many possible “roles” that a u-form can play at retrieval every distinct data object wherever
time is also described by its own u-form, which might point to
in the world its travels may take it.
many other u-forms.
“Toward the Universal Database: U-forms and the VIA
U-forms point to each other with absolute certainty by refer- Repository” —MAYA Design
ring to their universally unique identifiers. For example, a “col-
lection” u-form might have an attribute called “members”; the
value of that attribute would be a list of the UUIDs of all the u-forms in the collection.
Specific implementations of the VIA Repository can be run on centralized servers, and the
u-form-based storage architecture still provides the benefits of data liquidity and fusion.
During the evolution of VIA, MAYA has done just this under numerous corporate and
government contracts. This is also how many users, at first, will connect to the Information
Commons with conventional Web browsers, mobile phones, etc. These end-user devices
will not contain any Commons data or connect directly with each other. Everything will
happen through servers. In this transitional form or mode, the full vision of the Commons
is scaled back for compatibility with current technology.
But the real power of the VIA Repository comes when it is freed from the client-server model
and allowed to run in distributed form in a peer-to-peer (P2P) network. Thus, the Informa-
tion Commons will ultimately resemble a system closer to Napster or Gnutella than to the

18
World Wide Web. In fact, the visionary of the Commons, MAYA’s co-founder Peter Lucas,
sometimes whimsically calls it “Gnutella meets Encyclopedia Galactica.”
In this full P2P mode, the Information Commons will run on a universal, distributed da-
tabase that will be “everywhere and nowhere.” From the perspective of its users, this global
implementation of the VIA Repository will be “the great database in the sky.” They won’t
perceive (nor should they perceive) that its physical reality is in fact scattered throughout the
world among all the peer devices that access the Commons—including their own.
Shepherds and Replication

The permanent, unique identity of u-forms has the paradoxical effect of allowing them to be
in more than one place at the same time. If two u-forms have the same UUID, they are by
definition the same u-form—that is, identical instances of the same u-form. A “copy” of a u-
form, by contrast, has a different UUID than the original, and can go on to have a separate
life of its own.
Since no information device can ever have direct, immediate access to the entire Informa-
tion Commons, rules-based software agents called “shepherds” move replicates of u-forms to
and from network nodes when a u-form’s data is requested by a user. The more “in demand”
a given u-form is, the more it will be replicated throughout the system. Shepherds can be
built or configured by users of the Commons to reflect their information interests.
All u-forms can be annotated or commented upon by users, and unless a u-form
is specifically protected, users can modify the actual content of u-forms, too.
Because the system has this power, attentive readers will wonder how the replicates of a u-
form are kept in sync. If two replicates are changed at the same time, which one is authorita-
tive? If someone requests that u-form, which replicate gets shepherded?
There are a number of possible solutions, but one makes both social and technical sense. In
the Commons, u-forms have “owners.” An arbitrarily large number of other people might
make changes to a u-form, but those changes are not written to the u-form directly. Instead,
they are considered tentative “suggestions” that the owner of the u-form has the power to
“commit.”
A Global Information Economy

Financial economies that lack an abstracted, liquid currency are barter systems. You can ac-
complish rudimentary trade in such a system, but not sophisticated, ever-evolving exchange

that transcends the inherent meaning of traded objects such as silk or grain or livestock.
The idea of a liquid currency was a paradigm-shifting innovation in running an economic
system. All of MAYA’s thinking about information economy and device ecology is predi-
19
cated upon the very same idea, with the same revolutionary implications.
The common “currencies” of the bit, the byte, and the packet made massive and rapid evolu-
tion possible in computing and networking. In MAYA’s Information Device Architecture
(IDA), the basic unit of currency is the infotron, which makes
information devices truly fungible (mutually interchange- On the one hand, there is a clear
able). In the Visage Information Architecture (VIA), the
basic unit of currency is the u-form, which creates true data
need for some kind of ubiquitous
liquidity. All these concepts—currency, liquidity, and fungi- standardization on a grand scale.
bility—are quite deliberately borrowed from economics. On the other hand, the need for
What is the common currency of the Web world? Is it unbounded scalability places
HTML? XML? The Web “page”? The relational database stringent restrictions on the use of
table? The hard truth is that the Web has no common curren-
cy for information or information devices. As a platform for central authorities of any kind.
“the world’s information” (the phrase used in Google’s mis- “The Trillion-Node Network” —MAYA Design
sion statement), the Web resembles a comparatively primitive
barter system of “apples and oranges,” not a sophisticated economy.
Pervasive computing implies a true global information economy. Device fungibility and
data liquidity are its essential requirements—remarkably simple foundations for intense
complexity that remains comprehensible and useful at any scale.
Peer-to-Peer Is Not Evil

We will not be surprised if the mention of “peer-to-peer” has struck fear in the hearts of
some of our readers. If so, we hope it has not entirely closed their minds. The business world
has had some well-publicized experience with peer-to-peer networks—notably in the form
of the Napster file-sharing software and its ilk—and that experience is often portrayed as
traumatic (even if it really wasn’t). Because this type of file-sharing software was infamously
used to pirate copyrighted works, the entire category of P2P has come to be unjustly tarred
with the brush of “hacker crime.”
The entertainment industry in particular has gone after P2P with a vengeance, in some
cases lobbying lawmakers to try to outlaw P2P altogether. The latest example is the June,

20
2005 Supreme Court decision in the case of MGM vs. Grokster. But though this decision
was technically a victory for the plaintiffs, it was in no way a condemnation of P2P itself.
The decision simply said that if the owners of a P2P system actively encourage illegal behav-
ior, they can be held liable for that behavior. The Information Commons does not encour-
age illegal sharing, and in fact it makes extensive provisions for the protection of intellectual
property.
For the pervasive computing era, P2P networking makes vastly better sense than the client-
server model. Indeed, it’s the only viable alternative. MAYA’s leadership views P2P as essen-
tial to the realization of the true potential of Cyberspace—the ability to use information to
benefit humanity in ways not even imaginable today.
Intellectual Property and a New Information culture

Even though the blanket indictment of P2P by some business leaders has been wrongheaded
and reactionary, it is also true that no sane thinking about the future of information can
take place without allowing for the protection of intellectual property. Everyone, but espe-
cially commercial entities, will be relieved to hear that MAYA has given this subject intense
and detailed attention.
The Information Commons is presently populated primarily with publicly available data.
As its first proof-of-concept, it aims to enable a new era of civic computing. But proprietary
information intended for authorized participants or paying customers can and will co-exist
in the Commons alongside public information.
How will intellectual property be protected? Some technologists view P2P networking and
data security as fundamentally at odds with each other, perhaps even mutually exclusive.
MAYA strongly rejects this view, and has incorporated a number of mechanisms into the
Commons for accountability, authentication, and security.
First, the strong identity of every u-form (via UUIDs) makes it easier than ever before to
demonstrate diligence in protecting intellectual property, and to identify blatant viola-
tions.
Second, the intellectual property status of any bit of Commons data can be easily marked,
and at a fine level of granularity. For example, some u-forms in a dataset can be declared
to be fully in the public domain, while others in the very same dataset are declared to be
owned by their creators and to have some or all rights reserved. No other system today offers
a practical way to do this.

21
MAYA wishes to foster an information culture based upon a reasonable degree of sharing,
and so it encourages authors and publishers to adopt the Creative Commons designation
“some rights reserved” wherever possible, rather than compulsively marking everything “all
rights reserved.” But the Information Commons is by no means limited to the Creative
Commons philosophy (http://creativecommons.org).
Third, it is easy to digitally sign any u-form so that users can verify cryptographically that its
contents have not changed since it was signed by its publisher (the entity that vouches for the
information), even if it has subsequently passed through many P2P “hands.” Individual us-
ers of the Commons can also declare that they do not want data shepherded to their devices
unless those data have a certain intellectual property and/or signature status.
Fourth, the entire contents of individual u-forms can be protected via strong encryption
for those who wish to charge for the use of data, or to limit accessibility to privileged users.
Again, protected and unprotected u-forms can mingle in the same dataset.
MAYA Design welcomes inquiries from any organization, non-profit or commercial, that
wishes to explore participation in the Information Commons.
Humanservices.net: A Commons/Web Hybrid

In IDA and VIA, end-user devices are seen as “transducers” for receiving, processing, and
visualizing data. They are simply the means by which human beings experience something
that is inherently invisible to them: digital information. If information in the Commons
is really free to flow and fuse and manifest itself in any “transducer,” then the conventional
Web browser certainly qualifies.
This makes for a nuanced, subtle story in which MAYA is inventing a potential alternative
to the Web that can also play well with the Web itself. As we said earlier, the Information
Commons is a disruption that will succeed partly by making a graceful transition possible.
In fact, MAYA is undertaking a number of Commons-based initiatives, both commercial
and non-profit, designed to be experienced, at least initially, on the Web.
One such Commons/Web hybrid can be seen today in “preview” form at http://www.hu-
manservices.net/. This project, done for Pennsylvania’s Allegheny County, fuses county da-
tabases with those of the United Way charitable organization. Formerly, these two entities
maintained separate and largely redundant collections of data on human services available
in the same geographic area.

22
Besides the obvious waste of maintaining two sets of similar data, most of the information
was not even available to workers online because the agencies could not find a solution to
fusing their datasets. The humanservices.net initiative gave them not only fused data, but a
Web application that embodies some of the “information-centric” user-interface features of
MAYA’s sophisticated stand-alone software.
Humanservices.net: The Information Commons in a Conventional Web Browser
Some of MAYA’s “information-centric” user-interface features can be made available with the HTML and
JavaScript that conventional Web browsers understand.
As of this writing, the “preview” version of humanservices.net is intermittently available to

the public. The reason for this bears mentioning. The fusion of the two Allegheny County
datasets immediately made their discrepancies and errors visible in a way not possible before.
To avoid providing faulty information to members of the public who are often dealing with

23
personal crises, the agencies periodically take humanservices.net offline to fix the data prob-
lems that have surfaced.
All initial forays into data-fusion are likely to produce a similar experience. Fusion reveals
previously unperceived qualities of the underlying information—both positive and nega-
tive. No matter how sophisticated, information systems produce high-quality results only if
they have high-quality underlying data.
Civium Workbench: The First Information Commons Application

If the Information Commons can be experienced—at least to some degree—on the Web,
what happens when it’s not? How do you leave the client-server world behind, embrace P2P,
and leap wholeheartedly into MAYA’s great database in the sky?
To do that, you need to be running a peer-to-peer Internet application that connects you
to every other “user” of the Commons (which does not necessarily imply the existence of
a human being). Part of this application will be a software toolset for reading, visualizing,
and authoring Commons data. An equally important part of it will be a P2P version of the
VIA Repository installed on your own device (personal computer or otherwise). That’s what
makes you part of the great database.
Many such Commons applications are possible. In the future, they will come from a vari-
ety of developers. But the first one, Civium Workbench, is pure MAYA Design. Like the
browser-based application at humanservices.net, Workbench features an information-cen-
tric interface that lets users interact with data as directly as possible. Information-centric
interfaces get this power from u-forms, and are the result of many years of user-experience
and human-computer-interaction research at MAYA.
The Java-based Workbench, however, goes well beyond what can be done with the HTML
and JavaScript that Web browsers use. Notably, it gives users the power of “polymorphism”—
the ability to display liquid, fuseable data in many different ways at will, depending upon
the qualities available in the u-forms involved. Using Workbench, you might drag certain
bars of a bar chart onto a map to see them displayed geographically. (The bar chart and the
map simply display different attributes of the same underlying u-forms.) You might then
drag some of those mapped points onto a table, or a scatter-chart, or even onto a media-
player widget if those u-forms happen to possess audio or video attributes (or point to other
u-forms that do).

24
Civium Workbench allows data-visualization in space and time simultaneously. You might
manipulate an on-screen slider to portray on a map the relationship of toxic waste sites to
cancer incidence as it has changed over the last decade—watching the map update as you
slide through the years. Thanks to u-forms and VIA, the geospatial capabilities of Work-
bench exceed those of most commercial GIS applications available today. Some public agen-
cies have longstanding programs to collect photographic data on communities. With such
data imported into the Commons, you might, for instance, use Workbench to “drive” down
a certain street in 1990, 1995, or 2002, seeing exactly how it has been transformed.
Looking at Wildlife Preserves in Florida with Civium Workbench
Geospatial visualization is only one of many possible portrayals of data when using Civium Workbench
to “browse” the Information Commons. Note the horizontal slider beneath the map, which allows for
navigation in time. A public beta version of Workbench is expected by the end of 2005.
“Civium,” incidentally, is Latin for “of the people,” and refers to the flowering of civic com-
puting that MAYA hopes to engender with the Commons. One of core elements of the
company’s culture is the belief that it can be financially profitable while also serving the
public good. In the future, MAYA expects to launch non-profit and commercial ventures
that use the Civium brand.

25
Civium Workbench software and access to the Commons will be free of charge. A public
beta version of Civium Workbench is expected by the end of 2005.
Networking and “Emergent intelligence”

The dot-com mantra “the Internet changes everything” was wrong. It was wrong partly be-
cause it was not referring to the Internet itself, but to the World Wide Web, without ac-
knowledging the difference. The Web changed many things, but
hardly everything. And many of the changes the Web did make are Any organization scheme
simply not designed well enough to be permanent.
that requires each device to
But the mantra was also wrong in referring to the Internet, be-
cause that suggests an arbitrary limit on its significance. The phrase
receive any individual attention
should have been “networking changes everything.” whatsoever in order to join the
That’s a mantra worth chanting because it applies to the neurons of network is simply precluded.
the brain, or ants in an anthill, or human beings in a society, as well “The Trillion Node Network” —MAYA Design
as information devices connected to each other. The many “nodes”
of a network may not be very “smart” in themselves, but if they are networked in a way that
allows them to connect effortlessly and interoperate seamlessly, they begin to give rise to
complex, system-wide behavior that usually goes by the name “emergence.” That is, an en-
tirely new order of intelligence “emerges” from the system as a whole—an intelligence that
could not have been predicted by looking at any of the nodes individually. There’s a distinct
magic to emergence, but it happens only if the network’s nodes are free to share information
and processing power.
The realization of the Pervasive Internet will involve billions upon billions of protean net-
work nodes that ultimately “take on a life of their own.” Our present-day conception of
“intelligent devices” and global data networking does not allow for that. Until we change
that situation, we will not achieve the emergent magic implied by the phrase “the network
is the computer.”
All of this discussion—about the Web’s limitations and the great virtues of MAYA Design’s
big ideas—has really been about that.

26 Internet 0
A Zero-infrastructure ad-hoc network
The Pervasive Internet holds the potential for what Harbor Research calls “the networking
of every manufactured thing.” MAYA Design’s Information Commons incorporates the
necessary device and information architectures, but it also presupposes the existence of “a
zero-infrastructure, ad-hoc network” that makes seamless physical connection possible.10
How will the physical aspects of this networking actually take place?
Internet 0 is an emerging How will virtually any device, right down to a lowly light bulb, become
alternative [to other means a peer that connects at will to the global data network?
of device connectivity] that Obviously, billions of devices of wildly varying types cannot each re-
simplifies and unifies the ceive individual attention and configuration, or conform to elaborate a
physical side of Internet priori specifications. If it literally takes a network engineer to screw in a
smart light bulb, pervasive computing is never going to work.
access by using something like
Many schemes and so-called “standards” for device connectivity already
a Morse code for the Internet. exist. But of course, all those “solutions” add up to one big problem. We
—Neil Gershenfeld, Fab don’t want many standards; we want one.
An elegant answer is offered by the “Internet 0” (I0) initiative from

MIT’s Center for Bits and Atoms (CBA). I0 is predicated upon the idea that we already
have an excellent way to get every conceivable device on the Internet—the original design
principles and protocols of the Internet itself.
What the Internet Really Is

“Internet” is short for “internetworking.” The Internet was designed in the 1960s to allow
the incompatible data networks and computing systems of the time to share information—
to “talk to each other,” as people like to say. The Internet is literally a “network of networks.”
That’s what the “inter” means.11
The public Internet as we know it today is a worldwide embodiment of those original data
communications protocols—which are, by design, extremely simple. For this discussion,
Allmendinger and Lombreglia, “Four Strategies for the Age of Smart Services.”
10 Lucas, “Mobile Devices and Mobile Data—Issues of Identity and Reference,” p. 329.
11 Gershenfeld, Fab, p. 200.

27
their key attribute is that they make very few assumptions about the data they are sending
and about the devices connecting to the network to send and receive data.
It is this extensible, technology-neutral basis of the Internet that has allowed it to scale so
dramatically (and gracefully) since its inception, with minimal central administration.
Back to the Future

CBA’s name explains its mission—the convergence of the physi-
cal world of atoms and the computing world of bits. Like MAYA
There are twenty or so other
Design, it’s an inter-disciplinary organization; its members come
from many departments at MIT. CBA’s director, Neil Gershen- standards aimed, like Internet 0,
feld, teaches physics at MIT when he’s not embedding sensors in at networking embedded devices
Yo-Yo Ma’s cello or building “fab labs” where people can “print
out” such things as real working bicycles from their laptops.
rather than general-purpose
computers: CAN, LIM, BACnet,
Internet 0 is a back-to-the-future story. Just as the Internet’s
“internetworking” united disparate networks, I0 uses the same CEBus, LonWorks, X10, USB,
design principles for “interdevice internetworking”—uniting Bluetooth, SPI, I2C, etc. These all
disparate devices. Doing this requires a return to first principles. share one feature: incompatibility.
Thus the “zero.” I0’s name began as a whimsical response to the
—Neil Gershenfeld, Fab
high-speed “Internet 2” project, which aims to supplant the
present-day Internet (Internet 1) and enable ever more band-
width-intensive human-centric services such as IP-based television.
Internet 2 is all about speed and “big pipes” aimed at the senses of human beings. I0, by con-
trast, aims to create an “Internet of Things,” and things do not typically watch high-defini-
tion movies on demand. They don’t need lots of speed and bandwidth, but they do need to
connect effortlessly and interoperate seamlessly.
The Physics of Bits

To make most everyday objects connect and interoperate with the least engineering, high
speed is not only unnecessary, it’s bad. The reason is found in physics, something that most
IT professionals don’t think about very much. Bits have a physical size. The slower the speed
of transmission, the bigger the bit. Very fast, very small bits cover an entire building in a tiny
fraction of a second, and then scatter, requiring specialized network interfaces to capture
them. But a slower bit as big as the Ritz can embrace—physically—all the devices in the
building, from light bulbs and wall switches up to the commercial AC on the roof. When

28
a bit is that big, it transcends the need for specialized, expensive, incompatible network in-
frastructure.12
(And when we say “slow,” by the way, we don’t mean slow like a snail. At a data rate of a
megabit per second—the speed of most cable modems or DSL connections—a bit is about
300 meters long—big enough to encompass most
An Internet 0 Processor
buildings.)
I0’s creators (one of whom, Danny Cohen, was

a father of the Internet itself) liken it to Morse
Code, by which the very same message can be
blinked with a flashlight, pounded with a ham-
mer, or sent as electrical impulses. The trans-
mission medium doesn’t matter. Instead of the
“short” and “long” pulses of Morse Code, I0 uses
the 0s and 1s of an IP packet, but defines times
for representing them rather than the voltages
used by other serial transmission methods.
Just as the Internet uses the “end-to-end” prin-

ciple to remain agnostic about what is connected
to it, I0 extends that principle to “end-to-end
A full, working implementation of Internet Protocol exists on this
tiny processor. It is a self-standing, full-fledged Internet node. modulation.” Modulation refers to the physical
Source: MIT Center for Bits and Atoms representation of information. Like Morse Code,
I0 doesn’t care about the medium in which mes-
sages are sent.
Embedding the Internet into the World

I0 literally incorporates the Internet into physical infrastructure, making something like a
“smart building” much easier to contemplate than ever before. The building itself is literally
on the Internet and the Internet is in the very matter of the building. If this is such a good
idea, why hasn’t it been done before?
Much current IT practice in this area is a holdover from the early days of networking when
bandwidth was precious and implementing a TCP/IP “stack” in a small device involved
unthinkable overhead. That situation has changed, of course. Furthermore, I0’s creators re-
12 Gershenfeld, et al., “The Internet of Things,” p. 80.

29
alized that a lot of typical stack code exists to perform administrative functions that are not
strictly necessary in real-world, lightweight implementations like I0. “The code to run IP
can be squeezed into a few kilobytes and run on a one-dollar microcontroller.”13
As in MAYA’s Information Commons, I0-enabled objects live in a peer-to-peer model. Serv-

ers can add value to some I0 implementations, but they are not required by the architecture.
Like the “infotrons” at the heart of MAYA’s Information Device Archi-
tecture (IDA), I0 objects can be anything with information encoded in Internet 0 trades optimality
them, right down to a non-electronic house key. And like the u-forms of for simplicity and generality,
MAYA’s Visage Information Architecture (VIA), I0 objects maintain
their own identity with universally unique identifiers. which are as sorely needed
Each message sent by an I0 processor carries its own IP header. “The IP
in information technology
information adds about 100 bits to each message, which typically has as they are in the rest of the
a negligible impact on the response time and power requirements. In world.
return for this modest overhead, the network avoids the cost of config- —Neil Gershenfeld, Fab
uring and maintaining complex interfaces.”14
Internet 0 is a universal alternative to the many existing techniques for connecting ordinary
devices to the Internet. The other techniques all have something to recommend them; each
is optimized for a special purpose. But in return for their optimality, they sacrifice compat-
ibility. Since most device connectivity rarely requires maximum optimality, compatibility is
a much more important objective, and thus Internet 0 makes its trade-offs to that end.
An Internet 0 Packet in the Form of a Printed Bar Code
This is not a representation of an Internet 0 message, it is an I0 message. The vertical bars represent pulses.
Converted into electrical impulses by a barcode reader, they can be put on the Internet with no further translation.
The IP information adds only about 100 bits to the message, which, in this case, says “Hello.”
Source: MIT Center for Bits and Atoms


30 Conclusion
A Return to First Principles
The developers of the Information Commons and Internet 0 have not worked together, but
they have arrived at remarkably similar and complementary conclusions. In both cases, the
conscious insistence on a multi-disciplinary approach seems to have been a key ingredient in
allowing them to see past the fixed ideas of current computer science.
MAYA’s IDA and VIA, and CBA’s “Morse Code for the Internet” both succeed with a
“return to first principles” and by creating architectures that make the fewest possible as-
sumptions about devices and data and how they will be used in the future. Together, they
underscore the need to build simple, extensible foundations for intense complexity.
To many readers, these ideas will seem too good to be true. More to the point, they will seem
too simple to be good.
But that is the point. 3

References
Books
31
Dertouzos, Michael, The Unfinished Revolution, Harper, 2001.
Gershenfeld, Neil, When Things Start to Think, Henry Holt, 1999.
Gershenfeld, Neil, Fab: The Coming Revolution on Your Desktop—from Personal Computers
to Personal Fabrication, Basic Books, 2005.
Articles and Papers

Berners-Lee, Tim. “Web Design Issues: What a Semantic Web Can Represent.” Personal
notes on the Semantic Web (1998): http://www.w3.org/DesignIssues/RDFnot.html
Lucas, Peter. “The Trillion-Node Network.” MAYA Design paper (1999).
Norman, Andy and Lucas, Peter. “Information Architecture and the Emergent Properties
of Cyberspace.” MAYA Design paper (2000).
Lucas, Peter. “Mobile Devices and Mobile Data—Issues of Identity and Reference.” Human
Computer Interaction, Volume 16, pp. 323-336 (2001).
Lucas, Peter and Senn, Jeff. “Toward the Universal Database: U-forms and the VIA Reposi-
tory.” MAYA Design paper (2002).
Gershenfeld, Neil, Krikorian, Raffi and Cohen, Danny. “The Internet of Things.” Scientific
American, October, 2004, pp 76-81.
Knauer, Joshua. “The Next Generation Information Commons.” MAYA Design presenta-
tion (2005).
Allmendinger, Glen and Lombreglia, Ralph. “Four Strategies for the Age of Smart Servic-
es.” Harvard Business Review, October, 2005.
Further Information
The MAYA papers referenced above can be downloaded in PDF format from the company’s
Web site at: http://www.maya.com/web/infocommons/infocommons.mtml
The proceedings of an October 2004 conference on Internet 0, with links to supporting

papers, slides, and video, can be found at: http://cba.mit.edu/events/04.09.I0/

32 About Harbor Research, Inc.

Harbor Research Inc. has more than twenty years of experience providing strategic consult-
ing and research services to high technology clients. Harbor’s strategy and business develop-
ment work is organized around emergent and disruptive opportunities, with a unique focus
on the impact of the Pervasive Internet—the use of the Internet to accomplish global device
networking that will revolutionize business by unleashing entirely new modes of system
optimization, customer relationships, and service delivery.
Harbor Research’s clients are leaders in communications, computing, control, and content.
Harbor Research has built extended relationships with larger multi-line companies includ-
ing AT&T, ABB, Agilent, General Electric, Danaher, Eaton, Emerson, Hewlett Packard,
Hitachi, Honeywell, Hughes, IBM, Intel, Invensys, Lucent, Motorola, Rockwell, Siemens,
and Texas Instruments, as well as with growth companies such as EMC, Cadence Design,
Conexant and Qualcomm. We also work with a broad array of emergent start-ups and pre-
IPO technology ventures. We have built relationships with a number of significant Pervasive
Internet players, including Ember Corporation, Questra Corporation, Xsilogy, DataSweep,
and Dust Networks, to name a few.
Contact
Glen Allmendinger, President
Harbor Research, Inc.
gallmendinger@harborresearch.com
800.595.9368 ext. 24
415.615.9400 ext. 24 (outside U.S.)
fax: 415.615.0454

Designing The Future of Information: The Internet Beyond The Web

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Designing The Future of Information: The Internet Beyond The Web

Uploaded by

Copyright:

Available Formats

DESIGNING THE FUTURE OF INFORMATION

The Internet Beyond the Web

Current computer science operates with models of information,

Two unique initiatives—the “Information Commons” of

a harbor white paper

Harbor Research, Inc.

Designing the Future of Information 

“Internet Zero” (MIT’s Center for Bits and Atoms)

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

 hen it comes to preparing for the global

Pervasive computing—also commonly called “ubiquitous” or “invisible” computing—usu-

Lucas, “The Trillion-Node Network,” p. 3.

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

The “Information Commons” and “Internet zero”

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

 The “Information Commons” of MAYA Design

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

By virtue of MAYA’s unique information-architecture concepts, these datasets can be

Why Do We Need the Information Commons

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

“The Internet of Things,” p. 76.

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

Why Do We Need the Information Commons When We Have Google?

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

Data Fusion: Information Wants to Be Free

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

Isn’t the information commons the same as the Semantic Web?

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

“Infotrons” and Information device Architecture (IDA)

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

A Representation of “Infotrons” and MAYA’s Information Device Architecture (IDA)

The U-form: A Universal Data Container

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

A Representation of the Structure of U-Forms in MAYA’s VIA Repository

Lucas and Senn, “Toward the Universal Database,” p. 1.

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

In an unconnected or minimally connected world (a world of information islands), this

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

The Visage Information Architecture (VIA)

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

The VIA Repository: One Universal database

© 2005 Harbor Research, Inc. All rights reserved. http://harborresearch.com/

info@harborresearch.com • 800.595.9368 • 415.615.9400

Shepherds and Replication

A Global Information Economy

Designing the Future of Information

hen it comes to preparing for the global

The “Information Commons” of MAYA Design