IEEE Consumer Electronics Magazine - September/October 2019 - 52

Bits vs. Electronics

It allows me to transport (the information
about) the truck and, in effect, send the truck
through a thin wire.
We ignore the distinction in everyday language which creates a path-dependent understanding. If we ask the theoretic question about
channels, we accept the constraints as if they
were real. It does not help that the very words
we use make this easy-"communicate,"
"information," and "channel" all sound like their
day-to-day counterparts but are very different.
In fact, we go so far as to create channels out of
nothing when we allocate radio frequencies to
serve as dedicated channels.
In engineering, it is useful to build up abstractions so we do not have to solve all problems
from first principles, but we must not forget they
are just constructs that are meaningful only in
context.
The idea of sending a part number is not new.
A hundred years ago you could have called (or
telegrammed) me and told me that you had
admired the shirt I was wearing and ask me to go
to the store (say, a local Sears) and buy one just
like it to send to you. I would have told you that
you could go to the Sears near you and buy the
same shirt. Or, to order it from Sears, and they
would ship it from the nearest warehouse.
Yet, for communications engineers, if I pose
the question a little differently and ask how many
bits it would take to ship you a copy of an encyclopedia that takes a terabyte on disk you may
accept that I could compress it to a zip file so that
I only had to ship a few gigabytes. When I say I can
compress it to a few bytes (less than a 64-bit
word), I am told that, according to information
theory, it is impossible. Using an ISBN number as
my dictionary is not cheating any more than using
English words. Yet, communications engineers
implicitly use this very technique when sending
data-a shared clock is an out-of-band signal.
In the same way, stored programmed computing has evolved from high speed calculating
to something far more nuanced. Copernicus's
heliocentric solar system gave the impetus to
physics by showing a regularity that Newton
could build on.
There is value in thinking about names of
things vs. the thing itself. Naming is not a simple
concept. The distinctions we make between

identity and addresses are operational and not
fixed. One big idea is our ability to manipulate
naming or binding.
Which brings us to the idea that the truck in
New York and the truck in Seattle were the very
same truck. You may not have thought about it
that way because it was not a big issue at the
point because you know my nephew would not
care. But, it is a very big deal since so much policy and philosophy is based on a sharp distinction between direct and indirect actions. With
software there is not such a sharp distinction
and we have operationalized abstractions.
This gets confusing when we use existing
vocabulary in this new context. That is the way
language works-economy of mechanism. Once
we did depend on a telecommunications
industry which gave us the way to instantly communicate over a distance. Hence, we call it
telecommunications. Now, communications as
technology and communications as speech are
no longer aligned. Yet, in the US, we are left with
a Federal Communications Commission that is
betwixt and between. Other countries have their
counterparts, such as Ofcom in the UK.
I have written about how we discovered "best
efforts" by programming around an unreliable
middle using techniques such as retransmitting
packets from the end points. We also discovered
the power of better never than late and the idea
that we can fill in gaps rather than assume the
missing value is zero.
When we transmit a SKU number it has no
meaning to the facilities between the two end
points. That has another deep philosophical concept-the (meaning of the) data not existing in
the middle. The meaning only comes from context at the endpoints. This is even more interesting when we think of sending the toy truck across
the country. It does not exist in the between. And
the same process works for a full-scale truck too!
This concept of binding is central to programming. In 1958 John McCarthy came up with the
idea of a programming language based on
Lambda Calculus-LISP. LISP is used to operate
on data schemas rather than the data itself. This
is the reason it was an early favorite for AI. Schematic evaluation is also why relational databases
are so powerful. And those concepts are now
central to JavaScript.

IEEE Consumer Electronics Magazine

IEEE Consumer Electronics Magazine - September/October 2019

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - September/October 2019