A fork for the time-zone database?

Posted Sep 29, 2021 5:44 UTC (Wed) by flussence (guest, #85566)
Parent article: A fork for the time-zone database?

This seems like wildly inappropriate behaviour for what is supposed to be, and was previously defended in court to be, a machine-readable historic record of fact.

Imagine the chaos if Unicode's UCD was run this way and old codepoints got retconned to "tidy things up". +1 for forking this into the CLDR.

A fork for the time-zone database?

Posted Sep 30, 2021 11:41 UTC (Thu) by foom (subscriber, #14868) [Link]

Note that historic timezone data is certainly not immutable. It has frequently been (uncontroversially) updated to correct a mismatch between the database and some new evidence of the actual timezone rule in use at that time.

Furthermore, it has never pretended to be an actuate record of pre-1970s fact. That had been documented all along (e.g. https://data.iana.org/time-zones/tzdb-2016g/Theory)

A fork for the time-zone database?

Posted Oct 1, 2021 11:23 UTC (Fri) by tonyfinn (guest, #144891) [Link] (2 responses)

That would be an interesting discussion: "Levitating man in business suit" is too japan centric, better get rid of it.

Backticks are confusing because of dead keys and confusion with quotation marks, they only get into the valuable one byte characters because programmers designed the spec, better relocate them.

A fork for the time-zone database?

Posted Oct 2, 2021 1:48 UTC (Sat) by tialaramex (subscriber, #21167) [Link]

"Levitating man in business suit" is a symbol from an album by British ska band "The Specials", via the WingDings Windows font. So, nothing whatsoever to do with Japan.

A fork for the time-zone database?

Posted Oct 10, 2021 0:01 UTC (Sun) by flussence (guest, #85566) [Link]

Along those lines one of the weirder things Unicode does is to declare U+037E (Greek question mark) semantically identical to an ascii semicolon, not just a homoglyph or when NFK[CD] is used. A sufficiently pedantic text parser - which do exist in the wild - is allowed to silently replace the multibyte one with the single byte one. It's great for nasty source code tricks, but I'm not sure if it has a practical use.