Re: [irtf-discuss] Why the World Must Resist Calls to Undermine the Internet)

["[e-mail]" <[e-mail]>]

Trimming this to the more appropriate list...

> On Mar 16, 2022, at 1:20 PM, willi uebelherr <[e-mail]> wrote:
> 
> 
> Dear Hannes,
> 
> absolut correct, we act in the past and on the level of the past. In this time, the digital information, coded in analog signals, was going over a pure analog system with relais and amplifiers and on the end you have the digital information decoded.
> 
> In general, nature does not know digital technology.

You’re both overlooking quantum effects (by very definition of the word “quantum”), as well as many other effects that are countable but not infinitely so.

E.g., a single photon is an example of digital technology - it either arrives or does not.

Electron energy levels come in finite steps as well. So does single-slit self interference (it either happens or doesn’t).

The interaction of photons and polarized filters also exhibits binary properties (each photon either passes or does not).

Nevermind the system that gives us the word “digital” - fingers. Nobody counts in half fingers.

> Data transmission will always be analog.

Except when it’s digital, yes. ;-)

> Internally in our data processing systems we work with digital interpretation of analog signals.
> 
> In analog systems, no error checking can take place that relates to digital interpretations.

Analog systems have error checking too, just not the same way.

> But today, every node in the network works with digital information. Only the connections must still be analog. And thus the error checking can take place in each segment of the transport system.
> 
> TCP/IP is no longer needed. Also, because the resources in all end nodes are sufficient to receive the packets.

TCP reorders, recovers for loss, adjusts rates to account for congestion, AND maintains endpoint state. What do you propose will replace that and why would it not be reinventing it?

> If not, the transmitter can determine the capacity of the receiver beforehand, or the receiver will reject the reception.

This assumes receiver buffers are the issue; there are plenty of others.

Joe