[tz] Opaque tz names (was: Re: Reason for removal of several TZ)

[Philip Paeps]

On 2017-12-05 00:18:10 (+0100), Michael Douglass wrote:
> On 12/4/17 17:13, Clive D.W. Feather wrote:
>> Brian Inglis said:
>>>> I'd suggest a first step would be to provide a unique - reasonably 
>>>> short - opaque id for each definition. [...]
>>> Feel free to write a script to mv each path to its lat/long [...]
>> What's wrong with just labelling the present zones with randomly 
>> allocated letter-digit-letter codes? That gives us 6760 codes that 
>> don't look like city names or commonly-used abbreviations. We can 
>> then reserve letter-letter-digit for future expansion and 
>> digit-letter-letter for private use.
>>
>> As soon as people know it's lat/long, they'll try to map it to cities 
>> and the same arguments will start again.
>
> Exactly - that's why I said opaque

What problem are we trying to solve here?

Every couple of weeks someone asks on tz@ why their timezone is called 
City X instead of City Y or why City Z is not in the tzdb.  Every now 
and again that turns into a lengthy discussion on politics.  That 
problem is easily solved with a polite, terse reply pointing at the 
theory.

Whether we like it or not, there is necessarily a mapping (however 
opaque) between a time zone name and a geographical region.  We can give 
it any name we like but if we want to be able to keep the data accurate, 
we need to be able to map it back to the actual geographical region.

While the list of names in tzdb isn't great, it's better to have one 
list of seemingly arbitrary cities and one mailing list where the (let's 
be honest: not all that many) questions about them go.  Externalising 
this problem will only make maintaining the tzdb more difficult and 
needlessly confuse people.

I think I like the problem we have more than the new problems being 
proposed in this thread.

Philip

-- 
Philip Paeps
Senior Reality Engineer
Ministry of Information