[aprssig] IGATE message routing bug?
steve at dimse.com
Sun Nov 20 10:52:00 CST 2016
> On Nov 20, 2016, at 11:10 AM, Pete Loveall AE5PL Lists <hamlists at ametx.com> wrote:
> My software only uses hops but that is not necessarily the best way for everyone to do it nor is it the way everyone has implemented "local". But as I detailed earlier, the definition of "hops" is a fluid definition due to the multiple implementations over the years of various digipeating methods inclusive of the ambiguities in the "standard" used today.
The original meaning was clear, back in the days when the networking was simpler. Most IGates used traditional TNCs, where a single path was set in the TNC and not routinely changed. One could have a reasonable expectation that if, say, a station was heard by three hops it would not hear a message sent from an IGate with one hop in its path. While the advances have made that simple case not always applicable, the intent is still quite valid: an IGate should send out all messages that have a reasonable chance of being delivered with the path it will be transmitting those messages.
And that is where the sole reliance on reception of a position packet runs into trouble in my opinion. A mobile, for example, whose GPS has failed will not get messages. A guy who has mistyped his position won't get messages.
But even in those things on which I disagree I still recognize the wisdom of letting people choose. If someone feels a distance-based mechanism is best, they ought to be free to use it. But I would like to see the choice being mde by the local IGate operator rather than the IGate software author,
More information about the aprssig