APRS TIME STAMPS 12 Feb 2004 ---------------------------------------------------------------------- WB4APR APRS originally had time stamps in many formats, but the lack of consistency in PC clocks made on-air interpretation of received data ambiguous and full of errors. To solve this, in 1995, we changed all default position reports to NOT have time-stamps on transmission but for all software to add time stamps on receipt. RECEIVED TIME STAMPS: By assigning time stamps on receipt, then all packets heard are then self-consistent to the recepient's system and do give a live presentation of the data that is temporaly correct. ASSUMPTIONS: APRS is a real-time tactical communications system. Data moves at nearly the speed of light and so all data seen on RF must be assumed to be LIVE, or the most recent available data at that instant. As soon as the packet leaves the RF environment, a TIME STAMP must then and forever more be associated with that packet. LOG FILES: Thus all LOG files of ANY APRS packets must contain a RECEIVED time stamp so that this association with TIME is not lost on receipt. Note, some software saves RAW packets and no time stamps. This is NOT GOOD. Not only is the temporal nature of the entire log lost, but the file can then be reloaded or exchanged and the recepient has no indication of the difference between LIVE or OLD data. The only time an APRS packet can exist without a time stamp is when it is LIVE and on RF. TX TIME-STAMPS: Conversly, if anyone does use an APRS format that does include a time stamp, then that time stamp must be honored as the time when that data was valid, regardles of time of receipt. APRS INTERNET SYSTEM: Now the APRS INTERNET system has a problem. It also assumed that all packets are LIVE and so all IGates simply forwarded on LIVE packets and the time of receipt was the time of arrival at the other end of the APRS-IS. This then introduces many seconds or even ten's of seconds of delays. This cannot be solved by time-stamping on receipt by every IGate, because we have learned that we cannot trust individual PC times. APRS-IS TIME: I have no answer to this APRS-IS problem. Good luck. But I think it must all boil back to basics, and that is that at the instant that a packet leaves RF, that is when the time must be stamped for that particular copy of that particular packet, and that association must remain for all uses of that packet data... Conversly, on the slow 1200 baud channel, timing accuracies better than the nearest minute were never intended and this is why seconds were never included in the APRS formats. de WB4APR