[AusNOG] DSL Line Stats

[Ross Wheeler]

> And that's exactly my point: "Set a high SNR target figure" - SNR is the 
> RESULT. The "measure", not the "cause".

Just to throw another term into the mix (because it's relevant about 
now)...   NOISE.

If we define "Noise" as "any signal that isn't useful or desired", then 
there are many noise sources. Thermal noise. Noise induced from adjacent 
conductors. Impulse noise from car ignitions, mains switching transients, 
lightning, the plethora of electrical and electronic devices in use today.
Also, potentially, unwanted noise products from our own transmitters (both 
local and remote ends).

Under a given set of conditions, the noise is "relatively constant".
I can see how it could be argued that SNR is the "cause" of the speed, but 
it's really the NOISE that is the driver. If the NOISE were lower, the SNR 
improves (for a given level of signal). If the NOISE were to increase (and 
the signal remain the same) then SNR drops.

But SNR is still the "measure" and not the "cause".

(Or have I had it wrong since I passed my exams on this stuff back in the 
mid 1970s?)