[tz] [PROPOSED] Merge timezones that are alike since 1970

[Brian Inglis]

On 2021-05-29 14:09, John Hawkinson via tz wrote:
> Paul Eggert <eggert at cs.ucla.edu> wrote on Sat, 29 May 2021 at 14:43:31 EDT in <ba58913a-d230-ceed-3e57-2f3db21438d7 at cs.ucla.edu>:
> 
>>> merging zones across modern countries isn't acceptable.
>>
>> The practice is clearly acceptable, as tzdb's zones have crossed national
>> boundaries for decades in some cases (as I mentioned in an earlier email).
> 
> Paul: Words matter, details matter, language matters, and precision matters.
> Your statement is not correct, because you have blurred two different concepts.
> 
> There is a big difference between (1) "MERGING zones across modern countries" and (2) allowing to persist "zones [that] have crossed national boundaries for decades."
> 
> Taking an action that leads to a state is different from allowing that state to persist.
> 
> Thus, you may not fairly say that THE PRACTICE of MERGING is "clearly acceptable."
> (At least, not for the reason you offered.)
> 
> (Also, I think many would quibble with the claim that because the tzdb has done a thing, whether over complaint of the peanut gallery or not, that that thing is necessarily "acceptable." And with "acceptable" in question, I'm not sure what "clearly acceptable" means either.)

>> If we insisted on creating a new zone every time there was a different
>> country, we'd have a more-complicated database and get into even bigger
>> political messes than we already have. There was an example of this recently
>> in the complaint about why there is no separate tzdb entry for Kosovo. I
>> endured quite a bit of vitriol in private email about this.
> 
> Although this seems to be an oft-uttered sentiment, I am not entirely sure it were true.
> 
> If every political controversy results in a split pair of zones, I'm not sure that is so bad.
> If we had, say, Russia/Crimea and Ukraine/Crimea, the politics would be in which one to choose, but not in the maintenance of either.

More likely according to which is the largest municipality in each 
country's possession, and maybe something similar for E/Jerusalem.

> We would certainly have a database with *more entries*, i.e. a
> larger database, but I don't think it is fair to say that means
> "more complicated." Complexity is a deeper concept than just number
> of entries, and if we go from 600 zones to 900 zones or 1200 zones,
> I don't know that would have any meaningful effect on anything that 
> matters. I doubt the load on the maintainers would be greater, or 
> that a twofold increase would imply any changes to user interfaces,
> or anything like that.

>> In the long run we are better off decoupling tzdb entries from political
>> issues as much as we can. This will lessen the probability of similar
>> vitriol (or worse) in the future.

Given that the tzdb whole reason for being is the problems caused by 
political entities, the focus should always be on dealing pragmatically 
with all the political problems those cause, not following a course 
because it may be technically more pure, and not dropping all historical 
information because it may be worse, without good reasons.
Some(/many?/most?) downstreams maintain their own copies of older code 
and data and cherry pick updates to maintain maximal backward 
compatibility to avoid breaking their customers production systems, 
generating bug reports, that costs them effort, time, and money to deal 
with.
Organizations don't care what's right, only what's unnecessarily 
different: IBM, MS, Oracle, RedHat, and others have downstream products 
whose maintained lifecycles are decades, and they like to keep customers 
decades old applications running the same next month as last.

> But what does decoupling mean? Instead of fighting these political
> forces and maintain a steadfast refusal to acknowledge them and
> making people unhappy, why not just accept all the options and let
> the users choose? Wouldn't that be equally "decoupled"?
If we keep the rules for regions common, but have similar zones for 
different political entities at similar longitudes, we can maintain a 
balance, and still reduce effort.

-- 
Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

This email may be disturbing to some readers as it contains
too much technical detail. Reader discretion is advised.
[Data in binary units and prefixes, physical quantities in SI.]