Talk:GPS week number rollover

Suggested pagemove

Ping OhanaUnited and Patriccck. I just NewPagePatrol reviewed the article, made some tweaks for clarity, and added a link to this article from Global Positioning System.[1]

I would suggest that this article be moved from GPS Week Number Rollover 2019 to GPS Week Number Rollover with the necessary article tweaks, and a section for the 1999 rollover be added. Alsee (talk) 01:12, 10 June 2019 (UTC)[reply]

If someone will change introduction of article, you can move it. Patriccck (talk) 04:45, 10 June 2019 (UTC)[reply]

Done, and I've updated all pages that currently linked here/ Alsee (talk) 20:24, 10 June 2019 (UTC)[reply]

Questioning Revision as of 05:37, 1 November 2019

What do those Apple warnings have to do with the GPS rollover problem? WarpEnterprises (talk) 13:52, 16 January 2020 (UTC)[reply]

Suggested automated infobox

Something like what exists at the top of Swatch Internet Time or at French Republican Calendar#Current_date_and_time would be nice, where the reader is informed of ISO weeks since the launch of GPS (which is what the GPS week number would be if stored in as many digits as necessary for it not to roll over) and below it, the current GPS week number (so the previous number modulo 1024). Arlo James Barnes 19:20, 22 August 2020 (UTC)[reply]

I'll have to play around more but I left a commented-out framework in the article to allow people to see what I am thinking of. Arlo James Barnes 07:40, 26 September 2020 (UTC)[reply]

ISO week vs GPS week; time of event

The dates and time s of the rollovers that occurred already indicate

the rollovers occur as Saturday becomes Sunday
the ISO week changes as Sunday becomes Monday
the time scale for the rollover appears to be GPS time, not UTC.

These points should probably be added to the article.

Also there are comments in the article indicating an intent to add support for ISO weeks. The difference in days on which the two different weeks begin should be accounted for. Jc3s5h (talk) 12:23, 26 September 2020 (UTC)[reply]

That was fast! I just added those comments a couple hours ago. They should be regarded more as noodling around than a definitive plan for how to proceed. Indeed, the points you listed will all have to be resolved before and if my idea goes live, and probably more too. Unfortunately, we have to use the time-handling variables Mediawiki provides, which are largely ISO-based. But I think with some clever math it can be converted. Arlo James Barnes 12:30, 26 September 2020 (UTC)[reply]

Honda GPS

There was a GPS week rollover problem in some Honda GPS systems, where the clock would only say 0:00. Sometime later it started working again. The problem started, I believe, about August 18th 2017. I am not sure of the GPS week boundaries. Gah4 (talk) 08:06, 27 November 2020 (UTC) Checking on web reports, clocks started working again August 19th, 2019. Gah4 (talk) 08:19, 27 November 2020 (UTC)[reply]

I see that Honda (or whoever supplies GPS receivers to them) has a problem again. Otherwise, our 2002 Honda it still doing fine, though with about 1.5 years without a clock in 1017-2018. I didn't figure out the exact connection to GPS week, though. Gah4 (talk) 20:45, 7 January 2022 (UTC)[reply]

there definitely was *no* GPS week rollover in 2022: https://www.usno.navy.mil/USNO/time/gps/gps-week-number-rollover. The Honda issue most likely is still related to this, but "Honda was affected in January 2022, when GPS week 1023 rolled over" as the article states at the moment is wrong for sure. 2001:9E8:2B17:8000:3CE8:5B6D:28E9:A4CA (talk) 15:25, 8 January 2022 (UTC)[reply]

I've made some clarifying changes addressing this. 2001:9E8:2B17:8000:3CE8:5B6D:28E9:A4CA (talk) 15:33, 8 January 2022 (UTC)[reply]

Depending on how the data is processed, (and especially what mistakes are made) the effect can happen any time. Most likely on a week boundary, but maybe not even that. It might be 1023 weeks from when the code was written. Gah4 (talk) 22:50, 8 January 2022 (UTC)[reply]

Mention of Year 2038 Problem

Hi there - I made an edit to note that the Unix 2038 problem, which is described at Year 2038 problem, is a different problem to the GPS week rollover problem, as I felt that there was a risk of confusion between the two (both are timing related, both are binary rollover issues, both happen in 2038), and so it was worth noting that they were distinct. User talk:Strebe reverted this, but I'd like to propose that it might be worth keeping (maybe improving on my original wording). David Malone (talk) 15:40, 12 April 2022 (UTC)[reply]

Since no one has objected, I'll go ahead with this and add some text with an improvement on my original wording. David Malone (talk) 09:29, 1 February 2024 (UTC)[reply]