New home for time zone stuff by 2012?

Eric Muller emuller at adobe.com
Mon Aug 31 17:09:29 UTC 2009


Of the pieces of the puzzle that Arthur listed, some are not very 
problematic to transfer:

- data distribution
- code distribution
- mailing list maintenance
- mailing list hosting

I think everybody will agree that tz does not have special demands for 
those and that there are multiple adequate solutions. (In the interest 
of full disclosure, I am an officer of the Unicode Consortium, and I 
think that unicode.org would be a very nice solution.)

The interesting part is of course the rest. Before arguing one way or 
another, I think it would be a profitable exercise to try to 
characterize how the tz  project works. Here is my take; it is based on 
having followed this list for the past 5 years, but I am not a direct 
user of tz. I probably have some of the facts wrong, it's definitely not 
intentional.

Let's look at the data maintenance:

There are folks in the world who have a vested interest in having tz 
up-to-date. May be they incorporate it in their product, may be they 
live in countries with unpredictable DST changes. Together, they track 
the changes, and over time, they have figured out the best way: find 
some kind of verifiable source (if possible official) that describes the 
changes, and provide that to the list; it's even better if there is a 
patch that goes with it. Arthur collects those, and when he deems the 
collection large enough or there is something imminent, he proposes a 
cumulated patch. A little bit of checking, and bingo, we have a new 
version of tz. I seems to remember that at some point, Paul Eggert also 
played that role as much as Arthur, but I find less traces in the recent 
past.

I think the main reason this process works (and it works really well, as 
far as I can tell), is that everybody trusts Arthur and likes what he 
does: that he double checks the changes to ensure the integrity of the 
data (or at least publish the proposed changes), that he makes timely 
releases and that he does not take advantage of anybody. The last point 
is important: I have the sense that the community really believes that 
it owns tz, and that tz is nothing more than the community. There is no 
formal organization that owns it.

I have not followed the code maintenance as much, but I have the 
impression it follows about the same process. The needed changes are 
less obvious and take a bit more discussion, and the work is a bit more 
distributed.

So what is your perception of how the tz project works, and what aspects 
do you value (beyond the resulting database)?

Eric.




More information about the tz mailing list