Blog: 
Sep
Jun
Apr
Feb
Jan 2009
Your Device Hath Foretold
Dec
Nov
Sep
On the glowing screen of your mobile phone is a patent lie expressed
in vertical blocks. The image asserts a proclamation of strength.
Symbols indicate an estimate of remaining battery power, which in turn
promises more calls, but the skeptical consumer rightly questions
this claim. The bars speak falsehood. Your battery is about to die.
Your battery is fine. Also, I promise nothing will happen if you eat this tasty apple.
Not a Conspiracy
Although it’s tempting to suspect a secret alliance between
handset manufacturers and wireless networks, the real reason that
power meters are downright terrible has less to do with profiteering
and more to do with metaphysics. Some engineering tasks—such as
building a battery life indicator—demand innovation well beyond
the constraints of the natural universe. Although you can accurately
measure the amount of energy left in the rechargeable cells,
you cannot predict how quickly this reserve shall be consumed. The
future is the critical question, and to the product designers of the
world, the future is a frustratingly uncertain variable.
Not a Gas Gauge
One might reasonably assume that the power bars on a mobile phone
are analogous to the fuel indicator in the dashboard of your automobile.
A half tank means half empty, a quarter tank requests you take notice, and
a needle resting against the “E” practically barks
at the driver until they locate a gas station. The car performs equally
well regardless of the current quantity of hydrocarbons sloshing around
in its innards, so the fuel gauge is just a planning tool to prevent
becoming stranded. Shouldn’t your Blackberry do the same?
Unfortuantely, “E” does not mean “Enough.”
Unlike a personal vehicle, a cellphone is a device whose energy tanks
are tapped by many different parties. Only you have your own car
keys and can expend your own fuel, but anybody can call you
and run down your charge. Even if you don’t answer, the
phone must always be idling, ready to accept an incoming transmission and
chirp aloud the preselected obnoxious ring tone. Even if you are not popular
enough to warrant any inbound calls, the device silently performs a
continuous negotiation via its antenna, switching towers
as network traffic patterns change and as you physically move around town.
The battery drains, even though you aren’t actually on the phone.
Expecting a mobile device to correctly report how long the batteries
will last requires the device to somehow know how it will
be used in the minutes, hours and days to come. Handheld computers cannot
predict the future. The number of bars remaining is basically an educated
guess. This is why you can’t trust that meter.
Categorical Impossibility
The unrealistic demand for an accurate battery indicator represents
a transcendent burden of modern life. The science fiction classic
Star Trek explains it in terms of the relationship between
management and engineering. When asked for some unlikely feat from the
ship’s 23rd century technology, Mr. Scott replies in his Scottish
brogue: “Captain! Ye cannae change dee laws of physics!”
Time is a pesky dimension. Any attempt to predict or control the
future requires guesswork, and the farther ahead one attempts to gaze
the more impossible it is to conduct this endeavor with
any precision. We are so accustomed to absolute mastery over our
environment that we expect high reliability in every assertion.
Estimates based on the most unknowable of variables will be wrong,
but the technophile insists that every answer should be right.
Enter the Examples
Many common frustrations arise from an unwitting expectation
that machines can know their own destiny. It seems incredulous
that a computer endowed with the power to calculate a billion
operations per second and store billions of characters of text
would be unable to correctly foretell the amount of time
required to copy a file. Yet, until we invent soothsaying
circuitry, download progress meters will start out embarrassingly
wrong. File conversions, spreadsheet recalculations, database
exports, drive formats, backup jobs and software installation all require
some unknown quantity of time dependent on the changing
ecosystem of available resources. Nevertheless, consumers
would riot if the system answered sheepishly: “Time remaining
- Who knows?”
The problem is not limited the world of desktop and mobile computers.
Despite massive investment in fancy radar systems and monitoring
gauges, television forecasters cannot reliably call the temperature
more than a
few days out. Mechanics will usually
make an appointment to look at your car, but even then
they called their proposed service an “estimate.” Project
managers are always off on their schedules; financial managers are always
off on their budgets. The stars may well know what tomorrow holds, but they
certainly aren’t sharing it with us.
Need-to-Know-Basis
Specialization is the source of our technologically advanced society, and
not everyone who owns a mobile phone needs to understand the
electrochemistry of lithium-ion batteries or how frequency reuse factors
affect signal strength requirements. However, everyone should
understand how the nature of time affects the design and output of complex
systems. We all know individually that we cannot predict the future, but
we must also recognize that other people and the pantheon of man-made
technologies suffer from the same limitation. Your cellphone is
guessing about what might happen in the wild world of incoming calls
and undulating network traffic. Cursing at an unexpected drop in
power level is as foolish as blaming the dealer for a bad hand in poker.
Luck, whimsy and chaos are part of the game.
Although we cannot know exactly how the future will play out, our ability
to control and measure both current and upcoming factors allows designers
to continuously improve the quality of predictions. Take a plumber with
decades of experience to a leak he has personally diagnosed, and his
estimate of the work required will be impressively accurate. All predictions
start out terrible, but improve with collection of evidence and refinement of
the model. It’s reasonable to expect good answers about the past and the present.
For anything based on tomorrow, be ready for everything to be wrong.
Further Reading:
Would you like to leave a comment?
Read this.
###
If the heading above and graphic at right seem familiar,
it’s because they are recycled from
an 
While we can forgive cellphones for inaccurate battery meters,
false information in the signal strength meter and random dropped
calls cannot be excused under the same clause. Designing a network
with the density and capacity to handle everyone calling at the
same time from areas with the worst possible reception does
not require knowing the future, just a ridiculous amount of
money. Your battery meter is wrong because it’s not
Nostradamus. Your call was dropped because making a perfect
cellphone network costs too much.
