[tz] Call for evidence on UK adoption of EU summer time abolition

Paul Eggert eggert at cs.ucla.edu
Mon Aug 5 19:05:24 UTC 2019

Joseph Myers wrote:
> The House of Lords EU Internal Market Sub-Committee has published a call
> for written evidence on UK adoption of EU abolition of seasonal time
> changes.

Thanks for the heads-up. Attached is a draft comment to the sub-committee, in 
both .txt and .odt form (the sub-committee does not accept PDF). This mostly 
mimicks the comments I sent to the EU about a year ago. The main new wrinkle is 
the issue of naming and abbreviating time zones if Northern Ireland stays in 
sync with the Republic of Ireland while the rest of the UK goes its own way. 
Comments are welcome.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: UK-summertime-comment-1.odt
Type: application/vnd.oasis.opendocument.text
Size: 28681 bytes
Desc: not available
URL: <http://mm.icann.org/pipermail/tz/attachments/20190805/af6e18a1/UK-summertime-comment-1-0001.odt>
-------------- next part --------------
From: Paul Eggert
Organization: Computer Science Dept., University of California, Los Angeles
Email: eggert at cs.ucla.edu
Date: 2019-08-05

I am commenting as coordinator of the Time Zone Database (TZDB) hosted by the Internet Assigned Numbers Authority[1]. This database is used by most cell phones and many computers to track time zones and summertime arrangements. The comments below have been circulated on the TZDB mailing list and attempt to reflect consensus there. Most of these comments have to do with Question 7 (adequate lead time) of the House of Lords EU Internal Market Sub-Committee. The comments also touch on technical aspects of the change.

Although the simplest thing for computing applications would be to keep the current summertime rules unchanged, it would not be much of a problem to accommodate the abolition of twice-yearly clock changes in the UK, as has already been done in Russia and other countries. However, if such a change is done, the following technical points should be considered.

1. If any part of the UK changes timekeeping, there should be at least a year's notice before the change takes effect. Otherwise there is a good chance that some computer-based clocks will operate incorrectly after the change, due to delays in propagating updates to software and data. The shorter the notice, the more likely clock problems will arise. For more about this, please see "Changes to the tz database"[2] and "On the Timing of Time Zone Changes"[3].

2. If the UK rules are changed, the UK should record in a single public document all related changes to civil time, so that interested parties can easily track the timekeeping changes. For best results, the document should also describe the differences between civil time in the various parts of the UK, and civil time in EU and other related countries. This document could be purely informative, with no authority behind it.

3. If the UK abolishes twice-yearly clock changes, the resulting time zones should be considered standard time, not permanent summertime, to avoid conflicts with existing computer standards in widespread use. The notion of permanent summertime is not supported by the POSIX.1-2017 standard[4], which covers most of the computer servers in the UK.

4. Currently TZDB has a single entry 'Europe/London' for the UK. If Northern Ireland starts differing from the rest of the UK, TZDB will need to create a new entry 'Europe/Belfast' to record the UK time zone split. This will require more work for users in Northern Ireland who need accurate clock histories (e.g., for tracking financial transactions), as these users will need to change their TZDB setting from 'Europe/London' to 'Europe/Belfast'.

5. The UK should specify names and abbreviations recommended for any new timekeeping practices, to avoid naming confusion in computer applications. One such application is the Unicode Common Locale Data Repository, which is related to TZDB and which records time zone abbreviations and names in many languages[5]. For example, if Great Britain stays on GMT +01 all year, many English-language computer systems would call this "British Standard Time (BST)", following the historical practice for the UK from 1968 to 1971. If Northern Ireland aligns with the Republic of Ireland, it would be plausible for Northern Ireland to agree with the Republic of Ireland's names, which are "Irish Standard Time (IST)" for GMT +01 in summer, and "Greenwich Mean Time (GMT)" for GMT in winter; or Northern Ireland could use "British Standard Time (BST)" for GMT +01, agreeing with the rest of the UK in summer and making GMT its winter time; or it could use "British Summer Time (BST)" for GMT +01, the current practice in the UK. Government documents announcing the change should clearly state the preferred names and abbreviations, to avoid confusion during adoption.


[1] Time Zone Database, Internet Assigned Numbers Authority. 2019-07-01. <https://www.iana.org/time-zones>

[2] Changes to the tz database. 2019-07-01. <https://data.iana.org/time-zones/tz-link.html#changes>

[3] Johnson M. On the Timing of Time Zone Changes. 2016-04-23. <https://codeofmatt.com/2016/04/23/on-the-timing-of-time-zone-changes/>

[4] The Open Group Base Specifications Issue 7, 2018 edition, IEEE Std 1003.1-2017 (Revision of IEEE Std 1003.1-2008). See the variable "TZ" in section 8.3, "Other Environment Variables" <http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html#tag_08_03>.

[5] By-Type Chart: Timezones:Europe. CLDR version 33. Unicode Consortium. 2018-03-09. <https://www.unicode.org/cldr/charts/33/by_type/timezones.europe.html>.

More information about the tz mailing list