[aprssig] Fine Tuned New n-N Paradigm Filters

[Robert Bruninga]

>>> HamLists at ametx.com 3/28/05 1:00:38 PM >>>
>Someday, maybe digi owners will figure out the real way 
>to fix the mess on 144.39 is to get rid of source routing 
>completely.

Ah, But only if everyone uses APRS with only a single
one-and-only objective.  Yes, then the network can
"do what the user wants" because everyone by definition
wants what everyone else wants and the digi owner wants
too..  But APRS is used many different ways for many different 
uses and only the SENDER knows what he wants his 
packets to do.

Just look at the uses of APRS:  For WX, for Mobiles,
For INFO Queries, for DFing for storm chasing, for
local events, for state-wide nets, for internet messaging
for FINDU tracking...  All of these may need a DIFFERENT
path setting based on the SENDERS intent.

No one rule is going to fit even all the applicaitons of 
just one subset.  And I for sure do not want a digi to
be deciding how my packet should be routed based on
some SYSOP's decision without input from me.

> All the "New n-N Paradgim" has done is complicate 
>digi configurations, confuse travelers and digipeater 
>operators, 

I do not see how anyone can reach that conclusion:

The New n-N Paradigm "simplifies" from RELAY, WIDE, 
TRACE, TRACEn-N, SS  and any combination of the 
above down to one simple plan:  Use WIDEn-N where 
N is appropriate for the area.  How is that confusing?

>and cause new sources of duplicate packets (NOID off on 
>state/section routes) 

Again, quite the opposite.  There is no mechanism in the New 
n-N Paradigm that causes any additional dupes at all.  Setting 
ID in the SSn-N UIFLOOD parameter is so that routing is at 
least traceable to the last digi and so that people will *not* be 
tempted to use WIDE,SSn-N to get first-digi info but at the 
expense of more dupes (which was a lot of what was 
causing all the dupes in the first place)...

>while only temporarily reducing QRM in some places. 

I see the QRM reduction of DUPES and out of area packets
as permanent.   Droping RELAY,WIDE, TRACE, SS and 
TRACEn-N is permanent.  ALso the "perfect" dupe elimination 
of the WIDEn-N algorithm is not temporary and is permanent.
Thus the reduced QRM will only continue to be reduced as
more and more people finally start using the simpler and much
more efficient WIDEn-N (with small values of N).

> And this QRM reduction is many times at the cost of
>lost APRS stations.

I dont think so.  Even in LA we have seen an *improvement*,
tHe less QRM lets the smaller guys get in better.  Reduced 
QRM will always have an improvement in reliability with this 
system.  If any station is being "lost" under the New n-N 
paradigm, more than they were before, then his packets 
should not be being  sent in that area in  the first place.

>We have used systems without source routing for years 
>predating AX.25 (repeaters and linked repeaters ring a
>bell?).

Yes, but only when there is only one objective that works
for everyone, or when there is a higher level procedure 
for the sender to indicate his different requirements.  And
what could be simpler than for that user to chose N?
Or a direct path....?  In your case of linked repeaters, many
times you dont want to bring up the entire 5 state chain
of repeaters to talk between only two of the repeaters.
So the user used DTMF to configure the links.  And guess
what,  that is called *source* routing.... under the control 
of the source who is the only one who knows who he
wants to talk to.

Bob