[tz] Tzdb and the Sunshine Protection Act

Paul Eggert eggert at cs.ucla.edu
Thu Mar 2 22:44:49 UTC 2023


On 3/2/23 14:22, Brooks Harris via tz wrote:
> How will tzdb manage this?

Traditionally we've treated "permanent daylight saving" as standard 
time, and I'd rather continue this tradition than make an exception for 
the US. That is, tm_isdst would be 0. (Most people don't care about the 
tm_isdst flag, but POSIX and C standard nerds do.)

Whether the adjusted time in (say) New York would be abbreviated "EST" 
or "AST" or "EDT" is up to common practice. We could use the 
abbreviation "-04" until common practice settles down. If common 
practice becomes "ET" we couldn't use that, unfortunately, as POSIX 
requires at least three characters. At some point "EST" might become the 
best of the alternatives.

My biggest worry is the set of backward compatibility zones EST5EDT, 
CST6CDT, MST7MDT, PST8PDT as their continued use would lead to so much 
confusion that they'd be more trouble than they're worth. Presumably we 
would retire them by moving them to "backzone". "EST" and "MST" might 
need to retire as well. (Luckily, there is no "CST" or "PST".)

Similar issues will come up if EU regions go to "permanent daylight 
saving", as they have threatened to do for years.

Whatever we do in this area, it will be a mess.


More information about the tz mailing list