[aprssig] Balloon Paths again and again...

[Robert Bruninga]

> it sounds like you're saying that the WIDEn-N 
> digi's will see a packet digi'd by a WIDE1-1 
> as an original packet, even if they also heard 
> the packet directly, and digi'd it themselves.  

If the WIDE1-1 does just dumb digipeat, or does callsign
substitution, then yes.  That is correct.

> if so then this would be true also for any 
> terrestrial packets that happened to hit both 
> digis directly. 

Yes, this has always been the case.  And that is the primary
reason for abandoining RELAY and WIDE and going to WIDE1-1 and
WIDEn-N.  But stil each packet that actually hits a FILL-IN
WIDE1-1, still does generate extra dupes, but it is something we
live with, because the advantage of using WIDE1-1 fill-in's is
worth it in many cases...

> wouldn't it make sense to work towards smarter 
> dup checking algorithms in the WIDEn-N digis, 
> rather than complain when someone fails 
> to follow an exception in what was supposedly 
> a 'universal' path?

Many problems with those ideas:
1) The WIDEn-N digis all have perfect dupe elimination if
WIDEn-N paths are used.  
2) It is not the WIDEn-N digis that are causing the problem, but
the use of WIDE1-1 simple digis that do not decrement.
3) But the whole purpose of having a provision for WIDE1-1
fill-in digis is so that people can use any old TNC clone as a
WIDE1-1 fill-in, and not have to go buy a new KPC-3+ or other
new-N digi
4) Even the perfect fix to WIDEn-N digis to ignore the dupes
from simplistic WIDE1-1's would take 10 to 15 years to be
accomplisehd due to the large majority of absentee digis sysops.
And until all digis were converted, then you would still have
the same problems as now.

Sure, overnight, we could simply tell everyone with a WIDE1-1
fill in digi to go replace it.  That would completely solve the
problem (theoretically) but have no practical chance of success
for many years...

Bob