BG> How about these for shitty levels?  It was a 0011 call to the USR BBS

BG> I'd better look into the 0015 or 0019 prefixes again, 'cos these are fucked

BG> =================
BG> Freq     Level
BG>  150     39
BG>  300     29

PE> No, the USR SAYS they are fucked.  A completely different story.

BG> You might care to answer me just one question though - can you
BG> give me just one single scientifically valid explanation which
BG> proves that the USR's line probe frequency characteristics are wrong?

He isnt saying that, he is saying that there are some very
visible problems with the USR stats which mean we cant rely on
them until we know what the reason for the visible problems is.

I think that USR/USR stats are useful, coz there isnt any obvious
visible problem with THOSE, coz they correlate pretty well with
the other stuff like the cps thruput, but the clearly is a problem
when its not a USR at the other end coz EVERY set of stats without
a USR on the other end looks mad on the db levels.

That could be either because the USR stats are fucked by USR when
its not a USR on the other end, or because ALL those non USR modems
dont do the V34 line probe calibration stuff correctly. That later
isnt impossible, coz they are all rockwells, which is why the
stats when calling a Motorola would be interesting.

BG> You already know how they are derived, and at what point
BG> during the connect, so do your best to disprove them if
BG> you can. Funny thing is that US Robotics themsleves don't
BG> agree with yours and Rod's silly assertion either.

You've gone into zealot mode again, USR has said NOTHING about what
the cause of the visible problem with a non USR modem is due to.
Just done the usual finger pointing exercise, 'cant possibly be us'

AND I bet Frankowitz had never even noticed the problem either.
I bet if you had asked him if those stats would change just by
changing the V34 modem at the other end, he would have said no.
@EOT:

---