[EAS]The Mobile Internet

pjk pjk at design.eng.yale.edu
Sat Oct 13 19:59:05 EDT 2001

Subject:   The Mobile Internet

Dear Colleagues -


Another of those superb Economist technology surveys, this about
mobile uses of the Internet. 

And I append below a complementary essay by Phil Agre. At the
present stage of evolution of cell phone technology, he argues,
much good could come from unbundled cell phone capabilities
incorporated into wearable computing and other embedded roles. 

This resonates strongly with my own product design experiences. In
the late stages of a technology's evolution there are usually
various silly attempts at a fusion of functions, "product
convergence", that are usually duds in the marketplace.

What engineers at that stage are oblivious of is that they can no
longer design technology alone, but need to do it from an
understanding of the relationships of people and organizations for
which their technology is to be an enabler.

You might say that the design of advanced products _is_ tantamount
to the design of relationships. Technical virtuosity is necessary,
but no more sufficient than it would be for a Carnegie Hall

All best,  --Peter Kindlmann (aka PJK)

(from RRE News, 10/07/2001)

Blow up your cell phone.
by Phil Agre, UCLA

In the United States, much has been said about the dot-com meltdown.
In Europe, however, the salient meltdown is in wireless.  Government
regulators auctioned off spectrum for broadband wireless services,
and a combination of clever auction design and speculative mania drove
prices to insane levels.  Now large parts of the European wireless
industry is imploding.  The Japanese industry is doing better; at
least they, unlike the Europeans, have some proven applications for
their current-generation wireless services.  But the assumption that
people will dive into broadband wireless just because it's there is
not proving true.  Time and again, industry seems surprised by how
long it takes to establish critical mass for a new standard in the
market.  The Internet was misleading in this regard; its sudden growth
resulted partly from pent-up demand during the period of "acceptable
use" policies.  The point, in any case, is that the wireless industry
is driving off a cliff.

At the most basic level, the cell phone industry has lost the simplest
driver of innovation: reducing the size of the handset.  Cell phones
are now so small that, barring sharp turns in human evolution, they'd
be useless any smaller.  They can be made cheaper, of course, and
from different materials, but we are rapidly heading to a world where
cell phones as such are no more exciting an industry than calculators.
So where should cell phone design head next?  Part of the problem is
simply industry habit: except in Japan, whose experience does not seem
to generalize, cell phones have been used almost exclusively for two
applications, voice communications and text messaging, and the latter
application isn't even widely used in the United States.  The market
has been unified by a few standards and a lot of price competition.

The future will be different.  I'd like to see the whole concept of
a "cell phone" blow up.  A "cell phone" as we know it now is a bundle
of functionalities: microphone, speaker, buttons, display, internal
software, and various elements of the communications protocol between
the handset and the base station, among others.  Location-finding
functionality is on the way.  One direction of future development
already seems clear: instead of wiring the communications protocols
into the hardware, generalizing both the software and the protocols
with "software-defined radio" that can be changed dynamically.  But
another direction has had less attention: unbundling the functionality
of the cell phone and then embedding various subsets and supersets
of that functionality into a world of other devices.  Taken together,
these approaches -- software definition and unbundling-and-embedding
-- can lead to a vast new design space.

Here are some possibilities.  In Japan, I am told, radio stations give
out radios that consist of nothing but a credit-card-sized piece of
plastic with embedded electronics and a headphone jack.  The radio is
tuned permanently to the station that gave it out, so it doesn't need
a dial or display, and it's only meant to be used with the headphones,
so it doesn't need a speaker.  The same thing could be done with cell
phones.  Imagine a small object with a headphone jack and a single
button on it.  When you push the button, it "dials" a pre-programmed
number, such as a service that provides movie times.  If interaction
is needed then the button could be used, or else speech recognition.
If you add a microphone to the device then parents could buy it,
program it with their own number, and give it to their children.  It
would be like a specialized calling card, except that it would include
much of the functionality of the phone as well.

Unbundled cell phone functionality can also be embedded in personal
technologies.  If we imagine that we will all become cyborgs, carrying
around a mess of suitably streamlined gear, all of whose components
talk to one another, then the "cell phone" will surely need to talk
to personal sensors, databases, display screens, and so on.  These
personal technologies could communicate with other services over
the network.  This sort of thing has been explored by the wearable
commputing people.  What has been less explored is the main good
purpose for such services: maintaining awareness of the many people
and institutions with which we have ongoing relations: the kids at day
care, the public personae of our professional acquaintances, the ball
scores, the bus we hope to board, the discussion groups we monitor,
and so on.  (Traditional HCI research has drawn some lessons about
maintaining real-time awareness of work collaborators that presumably
carry over to wearable services.)

Another assumption of the industry has been that cell phones are for
mobility.  But from an unbundling-and-embedding perspective this need
not be true.  Imagine a historical battlefield.  At each point where
something important happened, the rangers have installed a green post
with a button and a speaker, and maybe a video display or other more
imaginative kinds of interactive devices.  The green post now has a
cell phone embedded in it, or certain functionalities of a cell phone.
The rangers use it to manage the interpretive "content" at a distance,
for example updating it when new scholarship becomes available or else
extending it with special features for significant historical dates,
material in other languages, and so on.  The posts could also include
public-address capabilities ("the park closes in half an hour") or
emergency call functions, etc.  In this case, the "phone" sits still,
being tied to a significant place, even as the people move around.
The posts could also interact with "augmented reality" gear that is
carried by the park visitors, for example by projecting diagrams or
animation onto the landscape.

The "cell phone" functionality could also be embedded in objects.
Warehouses already have a world of tracking technologies, such as
RFID tags embedded in the boxes, for keeping track of where particular
items are stored. (This is a serious problem for warehouse people, and
large objects get lost in warehouses all the time.)  With time, the
object tracking devices could converge with the unbundled-and-embedded
cell phone.  It would then be possible to pose much more general
queries to the objects, for example embedding sensors in the contents
of the boxes to assure their environmental conditions, run periodic
tests on stored electronics, etc.  Of course, when an object "phones
home" across an institutional boundary, for example between a consumer
who owns a washing machine and the manufacturer that sold it, the
relationship across that boundary becomes more complicated.  It will
be necessary to design the relationship along with the technology.

It is evident, I hope, that the design space of the the unbundled-and-
embedded cell phone is quite large.  It should be possible to look at
a particular application area and brainstorm a spectrum of possible
applications, each requiring different subsets or supersets of cell
phone functionalities.  That very diversity will pose a significant
challenge to the cell phone industry.  Look at the experience so far
with WAP.  WAP may well succeed -- its poor showing so far may simply
be another manifestation of industry's tendency to underestimate how
long it will take to establish a new standard in the market.  But WAP
itself and the WAP coalition display every warning sign.  Not enough
attention was paid to interface design, and the wireless industry did
not understand how to build the alliances needed to make the various
WAP applications really work as businesses.  They had lots of demos,
lots of start-ups, but little serious market acceptance.

What was needed, and missing, is a robust feedback loop between
applications experience and the basic design of the standards.
Useability problems became critical late in the day, rather than being
at the core of the design process.  Unlike the traditional cell phone
applications of voice and text messaging, a platform like WAP succeeds
only if it achieves economies of scale twice over: first in each of
a large number of applications domains, so as to make each of the
applications viable, and then in the applications taken as a whole, to
make WAP services viable in general, for example generating demand for
WAP-enabled handsets.

The unbundle-and-embed design paradigm makes the situation both easier
and harder.  On the one hand, if it really is possible to disassemble
the existing cell phone architecture and embed some of the components
into other systems, then that can only help the existing architecture
redouble its current economies of scale.  On the other hand, if that
unbundling-and-embedding strategy becomes economically central to the
industry, then it will surely place signficant pressures on the future
development of the architecture: the same problem as WAP, only worse.

As the trend toward embedded services unfolds, the design process
will have to change.  History suggests why.  The initial telephones
were fixed in place, either fastened to phone booths or tethered by
wires.  For many years one could speak of a person as "waiting by the
telephone" because the telephone was a place.  Cell phones changed
that, as phones become attached to people.  But the functionality
of the phone remained much the same, and the designer didn't need to
know much about the phone user's way of life.  As cell phones acquire
more features, more knowledge about users becomes necessary, and as
cell phone functionalities are unbundled and embedded, the resulting
services will become intertwined with the patterns of their users'
lives.  This, it seems to me, is the main line of development in the
history of communications services: a progressive intertwining between
communications services and the lives of the people who use them.

It follows that the design process of the future will require a
more sophisticated understanding of the user community.  This starts
with anthropological fieldwork, and it includes participatory design
processes, mock-ups and prototypes, and systematic mapping-out of
the whole universe of potential applications niches.  A good place
to start, as I've mentioned in the context of wearable devices,
is with relationships.  Think of the unbundled-and-disembedded
cell phone functionalities not as devices for making phone calls,
but as infrastructures for maintaining relationships.  What is the
informational architecture of a user's ongoing relationship with a
family member, a school, a doctor, a video game company, and so on,
and what could those architectures become?  What issues, privacy for
example, are at stake in the design and ongoing renegotiation of that
architecture?  How can the design process be part-and-parcel of the
large-scale cultural process by which people reimagine their lives and
choose once again the relationships that make them up?

More information about the EAS-INFO mailing list