You know, I have often wondered why some VWs and Audis have factory GPS navigation systems in them, but don't know the correct time! I'm not talking about high resolution either; +/- 15 seconds or so would be fine. GPS receivers do intrinsically know what time it is, don't they?
Yes, with accuracy on the order of nanoseconds within the receiver. I'm not deeply familiar with the technology, but my understanding is the time signal is compared between satellites to measure signal propagation delay and thereby triangulate your location, so incredibly accurate time is intrinsic to how it works.
And even if it's only UTC (or some equivalent), they also know where they're located, so how hard could it be for, say an system (that has over 100 MB of code!) to figure out the correct local time?
It's almost-UTC... it's off by a small but well-known number of seconds. UTC gets periodic leap-second adjustments and GPS time doesn't.
You do have to have a device with a little bit of CPU and the ability to get updates. Date and time math is shockingly non-trivial. A lot of developers think, what's the big deal, just +/- 3600 second blocks and you're good. Nope. Leap seconds, leap years, daylight savings changes, states and even individual areas within states that don't observe DST (Indiana in the past I think, also Arizona?), legislative tinkering (US in 2007, lots of other places recently), locales with timezone offsets that are sometimes 30 or even 15 minutes offset instead of 60 minutes, the
sheer size of the timezone rule data involved for a worldwide manufacturer, and the rate of change. This is something you do NOT want to brew yourself. You want to use the system or language libraries.
Edit: forgot to even mention the GPS geofencing for actually selecting a timezone before crunching all the data above.
The is well within range to do this. My wife's truck does it. My Phaeton syncs to GPS with a static user-set timezone, but it has the pre-2007 rules so it's worse than useless (this is what happens when you can't get updates!). A standalone low-cost analog clock display would probably not have the intelligence, but some of those (like my Phaeton at least) are dumb: they are commanded by the infotainment system to display whatever it wants, so they do okay. Doing this properly in say, low-line instrument clusters in cheap cars would probably be difficult.
Jason
