I've noticed that the press releases on the GPS rollover seem to lack a certain depth, so let's see if the programmers gathered here can add some. Apparently, someone is using the week counter to help compute the date. This means, of course, that the start date (Jan 6,1980) is hard-coded into these computer programs so that it can be used in the computations. In GPS receivers that are compliant, apparently the new start date of August 22, 1999 has been hard-coded, and a routine written that will tell the receiver to start using that date in computations. In GPS receivers that aren't compliant, however, come Sunday the date will be computed as Jan 6, 1980. If this is actually used to compute interest, etc. massive data corruptions will occur. It will also be interesting to see how long it takes to clean up from this and how long it takes to get a replacement for your GPS receiver once you realize it is obsolete. Do any other programmers have other guesses as to the internals?

-- Amy Leone (leoneamy@aol.com), August 20, 1999

Answers

To the top!

-- Amy Leone (leoneamy@aol.com), August 20, 1999.

OK, I give up. What does "to the top!" mean in this context?

-- so many expressions (so@little.time), August 20, 1999.

"To the top" means the answer was posted to move this thread to the top of the New Answers page.

The initial time point remains 23:59:47 UTC January 5, 1980. Not all receivers calculate from that, however: some calculate from the last verified time. Those that do calculate from the initial time point need to keep track of how many 1,024-week cycles have elapsed (so far, zero). Besides a week counter, there is also a second counter: the second counter is used for precision timing, not the week counter.

-- Lane Core Jr. (elcore@sgi.net), August 20, 1999.

To the top. Pretty clever. What I do to move my stuff "To the top" is argue with myself. Thanks for the tip.

-- Jimmy Bagga Doughnuts (jim1bets@worldnet.att.net), August 20, 1999.