> I claim that most of the time zone abbreviations that we have now are
> purely theoretical fabrications.
Hmm. How would you arrive at that conclusion? Most of the existing
entries appear in the zinc database and in other references, and I'll
assume that most of the places which these correspond to actually exist.
I've only been to some of them ;)
> What we need to think about is a system that is
> * customizable
> * internationalized
> * consistent
> * real
> The current system is none of these things.
Please give a comparison of "theoretical" vs the above. I'll agree that
extensible support would be nice, but would we be willing to look this
stuff up from system tables? If so, I'm happy to implement it. Anyone
have an opinion on what kind of hit we would see?
The same would be true, to some extent, for the internationalization
aspects of this, with of course the usual performance concerns about
cache misses and multiple lookups required to match tokens.
- Thomas