Hi James,

I am sorry but I do not understand the logic here:
"The ask for each province and territory in Canada to get a unique zone for the 
purpose of providing a unique name is out of scope for the tzdata project.”

The provinces and territories are larger than most countries and many cover 
more than one time zone in north america.  These are the jurisdictions that 
control time zones in Canada and have done since WWI. The timezones have not 
changed in 100 years, since the advent of computers.

We are simply asking tzdata to follow the law, so that we can follow the law.
This output from ‘date’ from my servers located in Port Alberni, BC, Canada, 
does not follow the law. Tzdata is the package that controls that output.

Thu 12 Nov 2026 08:16:37 AM MST

Please follow the law.

Cheers
Chris


> On Nov 11, 2026, at 04:15, James Bellaire via tz <[email protected]> wrote:
> 
> On 11/11/2026 1:31 AM, Brian Inglis via tz wrote:
>> The lack of effort by application vendors, and input from international 
>> contributors, are influencing the couple of volunteer maintainers to switch 
>> to just straight numeric UTC offsets (format %Z -> %z), in the hopes that 
>> applications will do their jobs better.
> 
> I did a quick count and there are approximately 439 "Zone" entries in tzdata, 
> including all regions and the back zones that exist for users/applications 
> who want historical time stamps. 173 zones use the "%z" format that passes 
> through an offset or a fixed numeric format such as "+05" that would provide 
> a similar result in areas that don't have DST. 132 zones have names that 
> change during DST (such as PST becoming PDT, CET becoming CEST or GMT 
> becoming BST). Note that due to the support of historic times there are many 
> zones that get limited use.
> 
> Support of accurate times since 1970 is within the scope of tzdata. I 
> appreciate the abbreviations being present. It is certainly easier when all 
> areas within a zone share a common name for their shared time zone. For 
> example, CET/CEST works nicely in Europe for Central European Time instead of 
> each country getting a unique name (which would require each country getting 
> their own zone in the main data).
> 
> The ask for each province and territory in Canada to get a unique zone for 
> the purpose of providing a unique name is out of scope for the tzdata 
> project. The focus is providing accurate times since 1970.
> 
> 
> I appreciate having names in tzdata and not %z. But areas that share a time 
> zone (such as America/Edmonton covering Alberta, eastern NWT and part of BC) 
> not agreeing on a common name for their zone leads to either the assignment 
> of an internationally known zone (such as CST which could be considered "well 
> known" as UTC-06 - unless you're referring to UTC+08 in China or UTC-05 in 
> Cuba). Or setting something less descriptive such as %z (-06 appearing in 
> applications relying solely on tzdata).
> 
> I will note that CLDR is passing daylight times for the recently changed 
> zones in Canada: PDT, MDT, CDT instead of MST, CST and EST. I would accept 
> that as a solution since those abbreviations have well known offsets (with 
> CDT being UTC-04 in Cuba being a minor confusion). There was some discussion 
> earlier this year about not wanting to use xDT names in zones where there is 
> no xST time to fall back to. The preference of using MST/CST/EST in areas 
> that have recently been PT/MT/CT vs following the CLDR example of using xDT 
> names without xST vs not providing a name (%z) is the choice being presented. 
> %z seems to be the least controversial.
> 

Reply via email to