[tz] Proposal to revert 2023b's Lebanon data changes

[Paul Eggert]

On 2023-03-28 08:53, Andrea Singletary via tz wrote:
> What did the airlines do?

As I understand it, airlines launched and landed planes as if the prime 
minister's change had not happened, and Beirut–Rafic Hariri 
International Airport posted two times for each flight, one using the 
prime minister's change, one without. Reuters reporter Timour Azhari 
tweeted about an airport time wormhole, passed along in The Register's 
Thomas Claburn's "Lebanon's IT folks face double trouble as leaders 
delayed Daylight Savings Time" 
<https://www.theregister.com/2023/03/28/lebanon_dst_delay_chaos/>:

> Azhari recounted passing through passport control at Beirut International Airport at 1200 GMT and, by virtue of inconsistent time keeping, boarding his flight 30 minutes prior to his security check, at 1130 BMT.


No matter what TZDB does, many uses on the ground will disagree and it's 
hard to judge which timekeeping is more popular. In today's "Lebanon 
fails the test of time" 
<https://www.ft.com/content/4788ed25-e925-4851-bf37-61699db5987a> 
Financial Times reporter Raya Jalabi wrote:

> “Our leaders have already broken everything in this country. I guess they got bored and decided to break time itself,” said Mohammad Sharif, a street vendor in Beirut, hawking corn.
> 
> Sharif said he would be keeping “Berri Mikati Time” (BMT, henceforth) as he’s fasting, unlike the frame shop owner next door, who said he would roll his clocks forward. “Lebanon is so democratic, every business can choose its own time zone,” Sharif said, echoing a joke that made the rounds on social media over the weekend. 


As I mentioned earlier there are good reasons to not split TZDB Zones in 
this case. So the choice here is between having 2023c simply revert to 
2023a (following one set of users and one interpretation of the law), or 
have 2023c reflect the prime minister's current spring-forward of March 
29/30 (following another set of users with a different legal 
interpretation - perhaps using "BMT" as an abbreviation? though this 
would need further discussion).

Reverting to 2023a has the virtue of simplicity for TZDB distributors 
and users, and where there's genuine uncertainty simpler is often 
better. Although this simplicity will inconvenience many users it seems 
like it is the best of a bunch of bad options.