[tz] Tzdb and the Sunshine Protection Act

Pete Resnick resnick at episteme.net
Sat Mar 4 21:12:59 UTC 2023

On 4 Mar 2023, at 11:46, Steffen Nurpmeso wrote:

> Eliot Lear wrote in
>  <867b8fc8-1229-dca7-8408-2c80818dfc6b at lear.ch>:
>  ...
>  |On 03.03.23 00:04, Tim Parenti via tz wrote:
>  |> On Thu, 2 Mar 2023 at 17:45, Paul Eggert via tz <tz at iana.org> 
> wrote:
>  |>     Whether the adjusted time in (say) New York would be 
> abbreviated \
>  |>     "EST"
>  |>     or "AST" or "EDT" is up to common practice.
>  ...
>  |> Worth considering that, if "EST" were to become standard for -04, 
> it
>  |> would require modifications to supported, but obsoleted, formats 
> in
>  |> RFC 2822 §4.3, which state:
>  |>
>  |> EDT is semantically equivalent to -0400

To be perfectly clear, even though the syntax was called "obsolete" in 
2822 and 5322, we are clarifying in the full Standard (being completed 
soon) that it would probably have been better to call it "legacy": The 
elements in that section are not going away in any meaningful way, but 
rather they exist in archives of email and are generated by legacy 
systems from time to time, as Steffen rightly (though perhaps 
indelicately) points out.

So as far as this syntactic form goes, "EDT" and the like should only 
exist in email archives or as generated by legacy systems (which no 
doubt will always believe that "EDT" is semantically equivalent to 
-0400), and therefore the meaning has not changed for purposes of 
interpreting extant email message "Date:" header fields. In current 
conformant email, those time zone designations are not generated.

>  |RFC 2822 is already obsolete, and this part was somewhat fixed in 
>  |5322.  See Section 3.3 of that document, in that those old zone 
> names
>  |are marked as obsolete, but allowed.  Pete can say more if he 
> wants.

There were no changes to this section between RFC 2822 and RFC 5322, and 
there won't be any changes when we move to full Standard. As above, when 
used in the interpretation of legacy email, those definitions are 
correct, independent of what happens in the future.

> Destructive comment, but i do not copy this.
> Unless i give in and accept that over fourty years of email
> soon disappear in the void of digital history, except for
> bitsavers or wayback machine(s).

A rather quarrelsome way to say it, but yes, in legacy email, these 
forms will still exist and still mean what they always meant.


Pete Resnick https://www.episteme.net/
All connections to the world are tenuous at best

More information about the tz mailing list