PE> The independent measure being another presumed-buggy USR?

RS> Which correlates well with actual thruput when the
RS> measurement is done with a Courier at BOTH ends of the line.

PE> You mean CPS rates?

Yes.

PE> What are you looking at to determine that?

Russ had some stats, particularly with the immaculate
FNQ/Bulldog and other sessions what werent so brilliant.

PE> Bill's files are far too small to get an accurate CPS reading?

Yeah. Russ has said he gets the 4000cps or close type CPS rates on
those phenomenal sessions he gets with the bulldog on decent sized files
that eliminate that small file problem, and its also the sessions which
have the phenomenally good frequency response curve, very flat over most
of its width. The poorer session like when calls Bill has much more slope
and the lower symbol rate correlating with the visibly worse graph.

PE> I reckon that's a tool binned for a long time.

RS> Yes, but you always do bin useful data.

PE> I bin suspect data, which can potentially do more harm than good.

Yes, but the real world is that when you dont actually have the
totally cast iron data, you may well be able to get useful info
out of less than cast iron data.

In this particular situation it would clearly be desirable to hire
some test equipment, and actually measure the FNQ/Bulldog sessions,
and the FNQ/Bill sessions, and the Bill/You sessions, and the FNQ/You
sessions and compare the results from that with what the Courier/Courier
frequency response curves claim the frequency response is.

When thats not practical, and the Courier/Courier stats do clearly show
that the FNQ/Bulldog sessions are MUCH better than say FNQ/Bill and still
better than FNQ/You, we do have SOME useful information on yours and Bills
line, particularly when the Courier/Courier stats do correlate with the
thruput in CPS. Its silly to just bin that data when you are short of data.
Tho important to remember that its a bit of a worry.

Thats why I would want to actually confirm that Courier/Sportster stats
are the SAME as Courier/Courier stats over the same session too.
@EOT:

---