[aprssig] APRSDepot.com

Steve Dimse steve at dimse.com
Mon Sep 11 08:20:56 EDT 2006


On Sep 11, 2006, at 7:35 AM, john b. leonard, jr. wrote:

> Outstanding work!  Thanks.  I put W9JBL-9 in there, and immediately  
> I could
> see where the digis didn't hear my sig.  I was using a 5 watt ht.

I must caution everyone about data derived from the APRS Internet  
System for propagation and network studies.

The APRS-IS was designed for the purpose of distributing the data  
payloads. Duplicate filtering removes packets at every point within  
the APRS IS (and to a limited extent on RF) to keep the bandwidth  
usage as low as possible. It  is estimated that only a tenth of the  
packets heard at IGates' TNCs are seen by clients on the APRS IS.  
None of the payload data is filtered, but perhaps 90% of the  
digipeater and IGate recipient data is filtered.

If the filtering were random, an argument could be made for treating  
the data as a sample, with error bars  determined by statistical  
means. Unfortunately, the data is VERY non-random, and non-random in  
a way that is dependent on impossible to quantitate  factors. At each  
point, the first packet to arrive will have its  entire packet   
forwarded. Two clients connected to different APRS IS hubs will see  
different digipeater and IGate info, dependent on digipeater delay  
times, RF propagation, internet propagation, and the exact  
interconnections present in the APRS Internet System.

Any data source which has 90 percent of its data removed in a non- 
random way cannot be depended upon. Use this data with care and with  
full awareness of its shortcomings. For example, the above use is  
fine, looking for places where a single transmitter was not heard,  
because this is based on the payload data. This same information is  
available on APRSworld, jfindU, and findU among others. On the other  
hand, using this site to see whether a particular digi or IGate  
covers an area is not an appropriate use, there are systematic  
reasons that a packet transmitted in a given position could always be  
seen on the APRS IS via a different digi or IGate.

Steve K4HG




More information about the aprssig mailing list