[aprssig] RE: New EU Paradigm also Uiview32 and New-N for the USA

[Ron Stordahl]

John

You said :"...Collisions would ensue because many TNCs will not wait and
listen if digipeating an alias, but they will if decrementing a
WIDEn-N."

I have never seen this...however all the area digipeaters are TNC2's 
with the UIDIGI-ROM firmware.  Perhaps this code does not have that 
issue?  Could it be unique to the KPC3's?  I wonder if this occurs in 
UIView32 digipeaters?

Ron, N5IN

John Habbinga wrote:
> Ron, when digipeaters are using WIDEn-N, then they will decrement by
> one, like WIDE3-3 would become WIDE3-2, then the next digipeater would
> decrement WIDE3-2 to be WIDE3-1, then the third digipeater would
> decrement WIDE3-1 to WIDE3* and the digipeating would end.
>
> If WIDE3-3, WIDE3-2 and WIDE3-1 were in the alias of all the
> digipeaters, then all the digipeaters would repeat the WIDE3-3 and
> stop.  Collisions would ensue because many TNCs will not wait and
> listen if digipeating an alias, but they will if decrementing a
> WIDEn-N.
>
> I have 4 digipeaters in my area.  We tried to make new paradigm work
> by using all aliases.  It didn't work.  Because there was no duplicate
> checking of digipeated aliases with Kantronics TNCs each digipeater
> repeated the same packet 3 or 4 times.  So if one guy came through and
> transmitted a beacon with WIDE3-3, then each digipeater would repeat
> it as many as 3 times, making 12 transmissions for every one beacon.
> We ended that experiment pretty quickly.
>
> So the answer for the NEWn-N paradigm isn't perfect, although it is
> much better than before.  At least you can minimize that aspect of the
> problem by not using WIDEn-1 as an alias.  Just let the digipeaters
> decrement normally.
>
> On 9/21/06, Ron Stordahl <ron.stordahl at digikey.com> wrote:
>> Cap
>>
>> The reply from John indicated that there might be some significant re
>> duplicate checking, which I indicated I didn't understand.  Perhaps I
>> should not have been to quick to make a change.  If you see no issue
>> here regarding duplicating checking I will admit I would rather have the
>> WideN-1 'literals' in the list as I initially did.
>>
>> Ron, N5IN