Re: Country specific add-on package

On Aug 26, 2004, at 4:23, Olivier Hallot wrote:
> There are 24 or so time zones.

Do you mean GMT-12 to GMT+12? No one cares about that, those are useful 
for almost nothing.
Timezones are a political thing and there are a lot more timezones:
   helge@power1$ find . -type f | sort|grep -v .svn|wc --lines
      120     120    1415
If I interpret this right, lF has about 120 registered timezones.

> All of them have a standard name. These names are already in place in 
> libfoundation. All of them MUST be there for a proper OGo operation as 
> you point.

Exactly. All political timezones must registered, because time 
calculations obviously requires the exact data, including DST info.

> But OGo/libfoundation is flexible enough to allow GMT-3 be named as  
> "GMT-03:00", Rio, Brasilia, Greenland, "Mom's TZ".

If this is correct, yes (but it doesn't make a lot of sense). But note 
that a timezone with abbreviation GMTxxx never has DST, because those 
are not political ones.
If Rio does DST, you either need to find out the timezone abbreviation 
or create a new timezone specification. In no case you can use GMT+03 
(because this is exactly that, Greenwich Mean Time + three hours, 
always ;-)

> User can choose one of these names inside OGo and he/she will be in 
> GMT-3.

No, this isn't something a user can choose. (he can, but he can also 
jump out of the window ;-)

> See, you *can* add a new name for MET such as Madgebourg.

Yes, but this doesn't make any sense.

> The webUI will show Madgebourg as timezone and that will be MET time.

How the WebUI localizes a timezone name is a completely different 
issues. If I want to show up MET as Magdeburg in a hypothetical 
Magdeburg.lproj, I need to provide proper labels, in no case the 
timezone spec would be changed for that.

> libfoundation is so good that you can also define "Madgebourg" and 
> "Madgebourg DST" in synch with MET and MET DST. No apointment missed.

Yes, because Magdeburg is just an alias for MET. Eg it would be very 
wrong to change the abbreviation to Magdeburg, it wouldn't work 
anymore.

> (Now what is not standadized at all is DST dates.

It _is_ - well standardized might be the wrong name, a name (eg CET, or 
MST) of a timezone (which _does include_ DST info) is registered in 
some database.
Of course daylight saving times are choosen by the government and even 
can be changed over the time - but even this is part of the timezone 
definition for eg CET.

> Some countries have them defined by astronomical rules and others no. 
> Brazil, for example has the DST defined by gov decree. Here we dont' 
> know yet if or when DST will begin in 2004, supposedly some sunday in 
> October.)

Ups, I begin to understand why you can't follow ;-)

If your country doesn't have a specified timezone, you need to 
workaround by using GMTxxx. I don't know why Brazil does that, but it 
results in inconvinience for the user.
The technically correct way to make calendaring systems work in this 
setup, is to make a _new_ timezone (America/Brazil) with _all_ the 
historical data. Eg OGo needs to know how it was all the years back, 
otherwise you can't properly browse into the past.
In addition - without DST information you can't create cyclic 
appointments which go into the future.

Again: a proper (political) timezone is strictly required! (whether or 
not you can predict them ;-)

Greets,
   Helge
-- 
http://docs.opengroupware.org/Members/helge/
OpenGroupware.org

-- 
OpenGroupware.org Discussion [email protected]
http://mail.opengroupware.org/mailman/listinfo/discuss