Re: Trimble Geoexplorer 1,2 date problems

On 28 Mar, 15:16, Terje Mathisen <"terje.mathisen at tmsw.no"> wrote:
Sam Wormley wrote:
Roman Szeremeta wrote:
Their excuse for producing incorrect data is that it doesn't matter!!!

  Because of the GPS Week Number Rollover, it is impossible with the
current
  navigation message to determine the correct date for more than 1024 weeks
  (19.6 years), with out an external date source or an internal clock
set by
  the user... no matter who is the manufacturer.
   http://www.navcen.uscg.gov/gps/geninfo/y2k/gpsweek.htm

That isn't quite true:

It is in fact trivial to check the current UTC-TAI offset and use that
to determine that the 10-bit week count must have rolled over.

You can even use the known offset at the last (first) rollover to check
if a second rollover must have occured.

This would allow the firmware to last for 70+ years from now.

That is usable, but I'm not sure you can rely on that for 70+ years as
that behaviour is quite unpredictable.
But anyway, wasn't it a patented method? It makes it basically
unusable (in this case... noone would pay for something that is not
used except for rare cases... I think)
.