runsforcelery wrote:The Safeholdian day begins with the first second of the first minute after Langhorne's Watch, which would be 00:00:01. Noon is 13:00:00 (for clock purposes), and most (not all) Safeholdian civilians start again with "1 in the afternoon" following that and end with midnight (or "13 at night") at 26:00:00. Langhorne's Watch isn't actually part of the day at all, as far as Safeholding timekeeping conventions are concerned, and as Weird Harold suggests, it is centered on celestial midnight so that celestial noon occurs when the sun is highest overhead, regardless of when "13 o'clock" happens to fall.
The Charisian military (and, increasingly, other militaries) use the equivalent of 24-hour timekeeping and count the first hour after noon as 14:00:00 and, as our own military, would call that "fourteen hundred hours." The civilians who use that timekeeping convention (which includes Charisians; one reason it was easy for their military to adopt it) would refer to that as "fourteen o'clock" or simply "fourteen" rather than "fourteen hundred hours".
As for how the Safeholdians build accurate chronometers, they do it by combining and cross-connecting two of them in a single case. One counts the hours and minutes of the official day and stops each day at the instant Langhorne’s Watch begins; the second counts the minutes of Langhorne’s Watch and displays them on an inset face. At the end of Langhorne’s Watch, the second chronometer stops and the first one resumes at 00:00:01. Trust me, it took them a long time to come up with a way that worked, and the . . . complexity of the challenge is one reason watchmakers have always earned a lot of money on Safehold. It’s also one of the reasons Howsmyn’s chief instrument maker was so good at his job.
Does that help?
For navigation purposes this complexity would be totally unnecessary, a single dial clock with a celestial time given as a solar angle (in degrees, minutes and seconds) would save a lot of useless complications in a mechanical time piece, with an optional a second scale for calibration in local time would be far more useful, if you want to add complications you could build in a version of the Antikythera mechanism and predict the position of a number of guide stars at a given time relative to your reference port and even optionally enter a guide star elevation and have it calculate latitude and longitude automatically. Having bells and whistles involved in interconnected time piece, or even clutch connected dual displays would be a recipe for loosing seconds per day due to the inherent mechanical slop in such a system, accuracy in a sea borne clock or chronometer requires attention to the basics of the time keeping mechanisms. Mechanical counter systems operating on similar principles are however not too hard to build and I have seen similar interrupted motion designs.
The regulator system in another matter, pendulums for example simply don't work in a moving reference system, unlike a land based clock, as a minimum a temperature compensated dual balance wheel system or similar novel system - this is totally unrelated to the complexity of the display. Issues like this is why hour glasses continued in use on ships centuries after accurate mechanical time pieces where in use on land.