Conversation

Notices

1. Glyph (glyph@mastodon.social)'s status on Thursday, 14-Dec-2023 07:53:33 JST Glyph

   I have worked on date/time systems https://xkcd.com/2867/

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Thursday, 14-Dec-2023 07:53:32 JST Foone🏳️‍⚧️

     in reply to

     @glyph you can imagine what happens when you're processing a few terabytes of digitized weather records and your datetime library runs smack into "02/29/1900"

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Thursday, 14-Dec-2023 07:53:33 JST Foone🏳️‍⚧️

     in reply to

     @glyph my favorite example of how date/time system rot your soul:

     When I worked for the government, we had a bunch of weather observations for February 29th, 1900.

     That's not a day that happened. 1900 wasn't a leap year. And yet, across hundreds of stations... people wrote down the weather for that day.

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Thursday, 14-Dec-2023 11:34:50 JST Foone🏳️‍⚧️

     in reply to

     • John-Mark Gurney

     @encthenet @glyph
     Sometimes.
     Sometimes they'd have February 29th but not March 1st. Sometimes they'd have both.

   • John-Mark Gurney (encthenet@flyovercountry.social)'s status on Thursday, 14-Dec-2023 11:34:51 JST John-Mark Gurney

     in reply to

     • Foone🏳️‍⚧️

     @foone
     Did those same weather stations also record the weather for March 1st? And was it different from Feb 29th?
     @glyph

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Friday, 15-Dec-2023 03:42:36 JST Foone🏳️‍⚧️

     in reply to

     • John-Mark Gurney
     • Tony Yarusso

     @TonyYarusso @encthenet @glyph so the answer turned out to be "it depends".
     Sometimes march 1st was missing because they realized the mistake before they sent in that (monthly) form, sometimes march 1st was the same as Feb 29, sometimes there'd be a gap later in the month. Sometimes there'd be no gap. Like, the station had years of uninterrupted daily observations, INCLUDING the day that didn't happen.

   • Tony Yarusso (tonyyarusso@infosec.exchange)'s status on Friday, 15-Dec-2023 03:42:37 JST Tony Yarusso

     in reply to

     • Foone🏳️‍⚧️
     • John-Mark Gurney

     @encthenet @glyph @foone I want to see some sort of analysis of “how long it took to realize their mistake”. Like, was everyone back on board by March 2nd, or are some stations off for months before having an awkward conversation with someone?

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Friday, 15-Dec-2023 03:44:06 JST Foone🏳️‍⚧️

     in reply to

     • John-Mark Gurney
     • Tony Yarusso

     @TonyYarusso @encthenet @glyph I was basically told "you can't fix it: leave it for a PhD to look at later"

   • Foone🏳️‍⚧️ (foone@digipres.club)'s status on Friday, 15-Dec-2023 03:45:41 JST Foone🏳️‍⚧️

     in reply to

     • Josh Dick :mac:

     @josh @glyph oh that's nasty

   • Josh Dick :mac: (josh@social.joshdick.net)'s status on Friday, 15-Dec-2023 03:45:54 JST Josh Dick :mac:

     in reply to

     • Foone🏳️‍⚧️

     @foone @glyph Mine was seeing American-formatted date strings being incorrectly consumed by a Spanish-language MSSQL test system, so that it would stop working after the 12th day of each month and “fix” itself on the 1st of the next month