Re: signals, communication, and bunnies

Luiz Ernesto Merkle (lmerkle who-is-at julian.uwo.ca)
Tue, 09 Feb 1999 09:47:20 -0600

Jay's and Bill's comments deserve more than I can afford writing
today. Therefore I'll limit myself to the first part of Jay's
comments.

> Luiz is quite right in adding some more precise engineering analyses of my
> signals examples.
I would say partially "quite write". :-O

If for some this will sound like a contradiction, it is a
contradiction. Technology is very context dependent, so much that the
community that have developed it usually abstracts away what is
around. Within its domain, one single wrong connection, and it
crashes. The concept of information has the same degree of dependence.
It is so confusing because in its formulation, it eliminates the
context.

I will illustrate this with a simplified description of how a
frequency analyzer works, what illustrates what Jay said about it. A
frequency analyzer is an equipment used to measure the range of
frequencies that a particular circuit responds to, among other things.
For example, it can measure how good are sound system, both the ones
people have in their head, and in their living room.

One of such an equipment, a digital one, is a system involving:

1) a spike or pulse of known duration and amplitude ( and the
frequencies it contains)
2) a equipment under test (the equalizer of a sound system, a
patient's hearing systems, and so on ...)
4) a frequency analyzer (the one that generated the pulse)
3) and a human,including a technician, physician, etc.
__________________________________________________

_ ___ ________ \\V//
| | <-> [___] <->[o o [ ]] <-> (O)~(O)
-- -- [________] oOO----J-----OOo
pulse filter analyzer human

___________________________________________________

Another way to measure the range of frequencies is to use a tone
(single frequency) as the input, and measure the output. Then change
the frequency, and measure it again, and so on. This will give the
same _information_ to the human, but it will be associated with a
different measuring practice, and a distinct set of equipment, a
different budget, ... In this context, a single spike has no
information, because it is too short for the associated equipment. It
cannot usually be measured. It is just noise.

> The spike does contain an infinite range of frequencies, but I really don't
> think it contains very much information, and I doubt you could really set
> many switches with one ... all the information would have to be coming from
> the analyzer.

I dare to say that the first use of word information is scaffolded
within the mathematical context, the second, a semiotic one. The first
is measured according to the possible forms that a sequence of pulses
can have. I agree, it has very little "information" in that context.
The second _information_ is connected to the uses that a spike can
have. In the above digital circuit, the pulse will be filtered by the
filter. Its output will switch a certain number of components in the
analyzer. The number of components will inform the human the response
of the filter.

The pulse itself, although has almost no information in the
mathematical sense, is key to information making in one practice of
measuring, but not in the other. Its absence eliminate the possibility
of having the information (digital analyzer) or just adds noise
(analogic procedure).

In either sense, the whole network, including the human and the
non-human, is necessary to achieve the _information_. All them, in
their heterogeneity, have to be connected and working properly. Even
the the mathematical concept has been developed in consonance
channels, filters, frequencies, tones, spikes, telephones, AM, FM,
TVS. They are in a unfinalizable dialogical relationship.

It is very easy to abstract the set events in a causal sequence. A
brief historical look in the models used in communication engineering
illustrate their change from a linear to an almost "fractal" topology.
Their material configuration though, even in the basic electrical
sense, need electrical currents going in both directions. The
recognition of that is important for accessing the use of information
"flow" within other disciplines such as semiotics, for example.

I would love to go further and talk about the difficulties that
teaching and learning the digital versus the analogic, the time
domain versus the frequency/phase one, and their relationships.
Nowadays, their relationship is even more messy, because technology is
a mixture of them. Communications engineers have big trouble to grasp
it and to switch back and forth according to the context embedded in
the technology they are developing.

That's it. I apologize in advance, but this spike of xmca
contribution postings from my part will become less and less frequent.
I know I'll be very tempted, but I'll be writing my thesis, which I
plan to finish it in a couple of months.

Tchau,
Luiz