[e2e] evolution of bandwidth as a term

[Cannara]

Panos, that's partly right.  "Baud" is a rate -- "symbols/sec", and what new
modulation tricks did was to add symbols that could be sent per second, so
increased Baud.  You've heard of FSK, QSK, etc. -- these are all
symbol-generating tricks that a bit-head wouldn't get, but an electrical
engineer who learned what channel properties (bandwidth, phase distortion...)
really meant would indeed understand and exploit.  Since each symbol added
allows another power of 2, if encoded with a bit, then the bit rates doubled
as Baud increased.

Understanding that communication, at any layer, is achieved by symbols (not
simply bits) is what leads to more of an understanding what Baudot, Nyquist,
Shannon and others did for us.  And, it should engender enough respect for
their work to stop misusing "bandwidth".

You're right that manufacturers led the way in misusing terms like bandwidth,
but such was avidly picked up by marketers in all areas, from chips to
systems, and not resisted effectively by those bearing scientific
responsibility -- us engineers.

It continues to amaze me how many people have a CS or other technical degree,
even write papers and go to conferences, and sit on committees (like
Kingfisher -- thanks JohnD :) and yet can't explain Baud, bandwidth, data
rate, or even throughput in one sentence.  We even have folks, in glassy
ignorance, creating new terms, to sound techy I guess, but to confuse new
people -- "goodput" comes to mind.  :]

Alex

Panos Gevros wrote:
> 
> I've heard people say that modem manufacturers are responsible for the confusion  - because as the bandwidth (usable freq range) on the physical medium stayed the same, new modulation schemes allowed the same baud rate (symbols/sec) to give higher information rate (bit/sec) but the manufacturers insisted on writting "baud rate = X bps" ( in the early modems there used to be 1 symbol = 1 bit)
> 
> Panos