wrote:

> > From: john_SPAM{at}wilkins.id.au (John Wilkins)
> > Thresholding is how an analog signal can become a
> > digital one
> 
> Yes. It's also how a noisy digital signal can be cleaned up.
> Feed the raw noisy signal into a high-gain amplifier, so that any
> signal even slightly away from the threshold voltage will produce
> output at one extreme or the other, only a signal almost exactly at the
> threshold value would produce any intermediate value of output. Combine
> that with positive feedback, and even if the output is initially
> intermediate it'll be further from the threshold voltage than the
> original signal, so it'll pull the circuit all the way to one of the
> endpoints, digital 0 or digital 1.
> 
> > The "trying" here is done by the sender and receiver,
and is imposed,
> > so to speak, on the analogue signal. What that "force"
can be I have
> > simply no idea. But I am not an engineer ...
> 
> If you don't understand how a simple electronic amplifier works, even
> at the black-box level (mathematical input->output function), if you
> don't even know what the word "volt" means in simple electricity, you
> need to study some elementary (high-school level) physics. If you know
> what "volt" means, and you know some simple electric circuit theory,
> like how to connect a battery to a light bulb so the light bulb goes
> bright, then here's a simple black-box explanation of an amplifier:

Thanks for telling me that in order to understand what "digital" means,
I need to go back to high school. I shall endeavor not to let my 20
years of using computers and systems analysis lead me into thinking I
know anything about the distinction. Particularly since it is a
philosophical distinction, rather than a physical one.
> 

> 
> Now in biological systems, if you understand anything about how enzymes
> work, you shouldn't have any problem understanding how an enzyme
> latches onto the reactants in particular relative positins, and how if
> the reactants are too far from correct position nothing happens, and
> how if the reactants are slightly mispostionned then a combination of
> electrostatic attraction and VanDerWals (sp??) forces will exert a
> force to align them better. On a linear polymer, there are only
> specific places where an enzyme can attach to the polymer, spaced once
> per monomer usually, so although the brownian motion that brings the
> polymer (such as DNA) and the enzyme (such as replicase) together is
> essentially an analog process, as the enzyme finds one or another
> attachment spot the forces convert the analog Brownian motion into
> digital attachment at specific descrete locations.

I also know that it is never that simple - these molecules do not always
pair up the way they "should", which leads to a distribution of data
when real world systems are measured. You are being misled by the
description/model. Naturally the texts will tell you the *usual*
process, but I know from talking to the researchers at my place of
meploy that they do not get "digital" results.
> 
> > you may appreciate this analogy.  The skilled bowler aims his
> > ball down the lanes using analog information.  The medium, (the
> > lane) is as flat as possible, to avoid distorting the bowler's
> > signal.  However, an unskilled bowler is producing digital
> > information.  The overwhelming majority of his balls end up in
> > the "gutter", and the gutters provide forces such that a variety
> > of inputs will mostly result in one of two possible outputs.
> > The ideal digital bowling lane would slope toward the gutters
> > from the center (the threshold).  And it is this "basin of
> > attraction" feature in the dynamics that distinguishes digital
> > from analog.
> 
> That's a great metaphor. The very long bowling lane acts as an
> amplifier, where a slight deviation from center line when the bowler
> lets go of the ball results in a correspondingly slight angular
> deviation which over the very long length of the bowling lane results
> in a much larger sideways displacement, causing the final sideways
> component of the ball to be often all the way to one side or the other,
> where a basin (attraction) generally keeps there rather than allowing
> it to bounce back to center of lane. Amplification can be increased by
> warping the alley to be higher in the center than along the edges, so
> that any ball slightly away from center will curve away from center
> causing the angular deviation to keep increasing.

Amplification is not what we were considering. We are considering the
digital nature of the outcome. If the ball hits a pin, though, the pin
may or may not fall, depending on how it hits it (angular rotation,
friction, etc), and in any case the ball hits analogue objects. The
"digital" nature of the hit is in the way we describe it (score).
> 
> > The point and logic of this example escape me entirely, I'm afraid.
> > And I do bowl. Some of my bowling makes it to the end of the lane, and
> > I even get a strike occasionally.
> 
> Do you understand how if the alley were very short, a given angular
> mistake would result in a lot less sideways distance deviation, making
> for hardly any gutter balls ever? Do you understand how if the alley,
> of normal length, were curved with a big hump in the middle, no human
> alive today would ever get anything except a gutter ball?

Spin bowling might become popular...
> 
> > But I was unaware of using digital technology or techniques to do so
> > - it was all analogue...
> 
> No, it's the mechanics of the alley, the hump in the middle and the
> gutters on each side, which apply an analog-to-digital process to your
> input. Your input was analog just fine. Do you understand now?

I understand that you are not understanding my claim, yes. *All* this is
analogue, in that there are no really discrete (macroscopic) processes
here.
-- 
Dr John Wilkins
john_SPAM{at}wilkins.id.au   http://wilkins.id.au
"Men mark it when they hit, but do not mark it when they miss" 
                                               - Francis Bacon
---
þ RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info{at}bbsworld.com

---
 * RIMEGate(tm)V10.2áÿ* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 6/21/04 6:10:52 AM