From: Warrl kyree Tale'sedrin (warrl@mail.blarg.net)
Date: Fri Mar 27 1998 - 14:01:33 MST
> From: Arjen Kamphuis <mountain@knoware.nl>
> I also read somewhere (but have not been able to confirm it) that the
> GPS-navigation satellites will 'run out of dates' by the end of _this_
> year. It seems their calender is based on counting weeks since 1-1-1980 and
> by next september or something they will run out of memory. Can anyone
> confirm this story?
I haven't heard it, but let's do some credibility analysis.
First analysis: is this way of tracking dates viable? It seems to me
that a GPS satellite would need a great deal higher precision than
this. If they used off-the-shelf date hardware of the sort routinely
used in PCs, then some GPS satellites may have a
year-2000 problem and some won't.
Second analysis: does this way of tracking dates lead to the
predicted effect? For convenience round to 50 weeks per year. A
one-byte week number would provide 5 years; this system, with 1980 as
the base, would run out of dates in late 1984, not 1998. A two-byte
week number would provide nearly 25 years, and run out of dates
sometime in 2004, not 1998. A three- or four-byte week number would
throw the die-off so far out in the future that the satellites will
be replaced before then anyway.
So the story as stated is clearly not credible.
(Caveat: sometimes *reality* is not credible.)
US$500 fee for receipt of unsolicited commercial email. USC 47.5.II.227
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:48:48 MST