[tz] Theory - proposal to delete the reference to population

[David Patte]

I believe a navigator, sailing to an island, would like to know its 
recomended timezone, without having to prove it has any people currently 
living on the island first.

On 2012-05-18 17:22, David Braverman wrote:
>> How is time offset observation enforced?
> The Time Zone Police, naturally. I think they're headquartered on Cape Cod.
>
>
> You know, having followed this thread intently for the last couple of days, I think we might be onto something. Clearly the tzinfo database needs to reflect the wall-clock time of every clock on the planet.
>
> Just glancing around my office, I count at least ten clocks, with offsets ranging from UTC+00:00:05 to UTC-05:02:10. Some of these clocks are on servers and use some approximation of UTC (it changes as the BIOS clocks lose or gain time over time), some are on laptops at UTC-05:something, and there's an antique clock permanently stuck at 7:03. (Or 19:03...I can't tell.) Of course, the servers and laptops occasionally reach out to NTP servers, so I'll have to capture exactly when they do this to make sure tzinfo maintains an accurate history. If someone would volunteer to go through all the system logs, we can probably piece together some of it, but unfortunately some of the logs have been overwritten.
>
> This logically requires substantial updates to the gazetteer. As I'm on the third floor of this building, we'll have to expand the locales/places/areas list to three dimensions. Oh--and my servers are stacked up five in a rack, so right there we've got five new time zones overlaying one square meter of Chicago.
>
> I almost forgot, I'm wearing a wristwatch, and I have a mobile phone, two digital cameras, and a handheld GPS receiver, all of which have (or are) clocks of some sort. We can ignore the GPS receiver, since GPS is pretty much the most accurate clock generally available. But still, I'm excited by the possibilities of creating a new zone for each of the mobile devices. Especially the wristwatches. We'll need to get everyone on the planet a notebook to write down when they adjust their watches, else the database just won't be complete.
>
> But how will we represent the watches/mobile phones/cameras moving through space? I'll be on an airplane tomorrow, traveling from the America/Chicago zone to America/Los_Angeles. I usually change my watch after takeoff. So we'll need to find some way to represent when that happens, because obviously that creates a new time zone right in seat 9F. I'll explain this to the flight attendants so all 150 of us can change at the same time, just to keep things simple. So look forward to an update tomorrow:
>
> + Rule  Flight581       2012 only       -       May     19 14:25        -2:00   F
> + Zone America/Chicago/Flight581        -7:00   -       PDT     2012 May 19 14:25
>
> Ah, but wait, I'm over-reaching. The key is "wall clock time." That rules out every clock not attached to a wall. It turns out, none of my clocks is attached to a wall, so never mind. It's still a vexing problem, though: how do we represent a time zone with no clocks in it?
>
> Wow. We've got a lot of work to do. Let's to it, then.
>
> Or, in the alternative, we can just do our best to capture the general rules for approximate areas of the planet's surface that have both people and clocks, and call it "good enough."
>
>


--