The real question is, *where* in the system - at the individual
Aggregators (1 per Skimmer), at the RBN database server, at the RBN
telnet server(s), or at the user's local Telnet server? I tend to agree
that we need a centralized solution, preferably one that makes both the
unprocessed stream and a processed stream available for users who may
want one or the other.
But before we can do any of this, we want to test CT1BOH's algorithms,
both for accuracy (no fair rejecting real, good spots) and for speed.
Any solution must be capable of at least 40 spots/second *average*,
preferably more. We've only been looking at this for 48 hours, and
everyone on the RBN team except me still has a real job, so it won't
happen overnight, but we *are* serious about it.
73, Pete N4ZR
Check out the Reverse Beacon Network at
http://reversebeacon.net,
blog at reversebeacon.blogspot.com.
For spots, please go to your favorite
ARC V6 or VE7CC DX cluster node.
On 7/17/2013 4:47 PM, w5ov@w5ov.com wrote:
Jack,
That would be OK if the busted spots would stop coming in. Think of the
wasted bandwidth around the globe that results from this problem. Adding
to the processing burden of my local logging program is not the answer
when it can be done in the RBN system itself and totally eliminate the
problem.
You have to remember a couple of things:
1) is that these are busted callsigns. They are copied wrong by the
skimmer systems. Just because they happen to be wrong, does not mean that
they might not also be a real callsign so a "blacklist" will not work.
2) is that the RBN is not like the human packet network where you could
send a "TALK" message to someone sending out busted spots and tell them to
stop. The RBN cannot be told to stop making mistakes - and it
continuously repeats the same errors over and over and over and over and
over and over and over and over and over and over and over and over and
over and over and over and over and over and over and over and over.
Get the point?
Error correction in the RBN is the correct solution.
73,
Bob W5OV
Me too - another casual contester.
A simple enhancement of the logging program(s) you use might help a lot
with those serious operators' problem. Suppose you could have a command
to
tell your logger "treat this call as if I worked it", so it would never
appear again as a spot on your screen.
In N1MM, for example, there is a "remove selected spot" menu item now to
get rid of a spot on the bandmap. A similar item "Remove selected spot
and
never consider that callsign again" would stop the "over and over and over
again" problem.
The loggers could even have a file, similar to the ones we use today as a
"whitelist" of known contest calls, but containing a "blacklist" of known
bad calls. The loggers would simply read such a file and then ignore any
spots for those calls.
It's good to always improve the accuracy of the data feed, but there are
other techniques that might be very helpful too...
73,
/Jack de K3FIV
On Wed, Jul 17, 2013 at 5:10 AM, Bob Naumann <W5OV@w5ov.com> wrote:
For the more "casual" contest operator (of which I'm one - most of the
time), and for the day-to-day DXer, the errors admittedly don't present
much
of a problem.
The problem comes in a major DX contest, when you're a serious operator,
and
during the last 24 to 36 hours of the contest you've already worked most
everything that is unique (multiplier-wise) on the band(s). As a
result,
your logging program is no longer alerting you to the stuff you've
already
worked (the real callsigns). Guess what you do see? A preponderance of
RBN
errors! It has been at a level that is simply unacceptable. Most of
these
are not "off by one" errors either. The real problem is that they never
stop. The same errors are propagated over and over and over and over
again.
It becomes quite unsettling after a while. "When will this stop?"
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|