[e2e] Re: evolution of bandwidth as a term

Cannara cannara at attglobal.net
Fri Oct 3 22:12:07 PDT 2003


Exactly, Nitin.  "digital bandwidth" means what?  The true bandwidth of the
channel(s) carrying the digital signalling?  The bandwdith of each line in a
bus x the number of lines?  Bits/bytes per second?  What, indeed?  Using the
"digital" prefix adds nothing but further confusion.  The whole point in the
word is that it's extremely well defined in electrical engineering, signal
theory and physics.  Just because some folks decided to use it incorrectly
means nothing as far as now suggesting it needs changing.

And, as you say, "information rate" is indeed in the semantics of the symbols
received and the receiver's interpretation of them, just as beauty is in the
eyes of the beholder.  The fact that symbols can be recoded in some number
base has nothing to do with bits/second, trits/second, qubits/second, or
whatever.  Information rate, as defined by Shannon, used binary encoding as a
simple representation, not as a mathematical or physical necessity.

The bottom line is we have plenty of proper terms and there's no logical
reason to avoid them.

Alex

"Nitin H. Vaidya" wrote:
> 
> My first attempt at this posting apparently got held up because
> the filter found my e-mail header "suspicious".
> Trying again with alternate subject ...
> 
>   >>Information rate is a function of the semantics of the symbols; in the
>   >>absence of that semantics, information rate = bit rate.
> 
> Well, may be I am misinterpreting the statement above, but the
> way I understand "infomation", infomation rate is NOT bit rate,
> semantics or not. Bit rate of a data stream, and the rate of
> communicated "information" are often unequal (as someone recently
> illustrated with the example of CNN).
> 
>   >>Perhaps it's sufficient to use a qualifier?
>   >>
>   >>    analog bandwidth
>   >>    digital bandwidth
> 
> Sufficient perhaps, but is it necessary?
> Besides, "digital bandwidth" still leaves the door open for
> a confusion between "bit rate" and "capacity", those two being
> different. Which one does "digital bandwidth" correspond to?
> 
> Wouldn't it be far easier to fall back on "bit rate"
>X-Mozilla-Status: 0009at is what one means?
> 
> - nitin





More information about the end2end-interest mailing list