Oh… it's 7/7. That means it's one of just twelve days a year I know what the date is in Canada

@mcc do they have different days 🤔

@aeva so in America they write the date like MM/DD/YY and in Europe they write the date like DD/MM/YY and in Canada they just sorta… wing it. Pick an order. Whatever

@mcc may I propose a compromise: MD/DM/YY

@aeva @mcc Mid-endian dates? noooooooooo!

Reminds me of how Teledyne LeCroy oscilloscopes timestamp waveforms in nanoseconds since midnight January 1st, 2000...

... In the instrument's current time zone.

This means that if I pull a waveform off it right now, the timestamp epoch is Jan 1 2000 at midnight *Pacific Daylight Time*. Oh, there's no DST in January? Too bad, there is in LeCroy scopes.

@azonenberg @aeva @mcc ... wow that's *horrid*. I bet just running the scope set to UTC is annoying, too...

@astraleureka @aeva @mcc Well the bigger issue is that as a piece of software connecting to the scope (or loading a .trc waveform file using the same header format the scope uses for live data) it is impossible to know the actual trigger time without external information (the scope time zone).

You might personally set your scope to UTC, but software talking to it has no way of knowing that.

@azonenberg @astraleureka @aeva one of the big moments for me in losing Trust in computers was early in my programming career when I had two devices connected via a serial port, and despite both devices being synchronized to atomic clocks (one was running NNTP, the other had onboard GPS) no matter what we did they continuously disagreed on the current time by *exactly* 14 seconds, which was causing bugs.

Do you see?

Do you see the problem?

@mcc @astraleureka @aeva Disagreement over leap seconds? Long RS232 packets at low baud rate that took 14 seconds to exchange data?

@azonenberg @astraleureka @aeva Disagreement over leap seconds. GPS time forked from UTC in 1980 :(

@mcc @azonenberg @astraleureka what solution did you end up going with for this problem?

@aeva @azonenberg @astraleureka Either we had the two systems exchange current time at boot and calculate an offset or we hardcoded 14 seconds. I hope not the latter because the company later ceased to exist and would not have been able to ship updates the next time a leap second was announced.