RE: [obix-xml] TimeZone issues

["Brian Frank" <bfrank@tridium.com>]

1) I think I prefer “tz” myself

 

2) I definitely lean toward city name only - it is trivial to populate a lookup hashmap with the city name once you have the database of zones.  And I personally find it much nicer to use the city name only:

 

  val=”2007-11-13T15:45:00-05:00” tz=”New_York”

  val=”2007-11-13T15:45:00-05:00” tz=”America/New_York”

 

3) I think you are suggesting that the client timezone be the default?  That doesn’t make sense to me – if anything I would think the default would always be the server’s timezone.  But then again I’m thinking about oBIX in general where most times are associated with things like histories and alarms (not necessarily traditional meeting scheduling).

 

---

From: Richards, Dave [mailto:drichards@trane.com]
Sent: Tuesday, November 13, 2007 4:41 PM
To: obix-xml@lists.oasis-open.org
Subject: RE: [obix-xml] TimeZone issues

 

1)       I like “tz”.

2)       Brevity would say use the city, but since the values are already defined elsewhere, I’d vote for keeping exact values.

3)       I like (a), and if not defined it is assumed to be local time of the receiver.

 

 

Dave

 

---

From: Brian Frank [mailto:bfrank@tridium.com]
Sent: Monday, November 12, 2007 2:08 PM
To: obix-xml@lists.oasis-open.org
Subject: [obix-xml] TimeZone issues

 

I haven’t gotten a chance to work on the actual document.  But I do have a list of issues I’d like to make a decision on for Wed’s meeting.  Previously we decided that timezones would be defined via a facet with a value space of an zoneinfo/Olsen database identifier – the Olsen database itself would define the timezone and daylight saving times rules for a specific identifier.

 

1) Should the facet attribute name be “timezone” or “tz”?

 

2) The Olsen database uses timezone identifiers such as “America/New_York”.  The continent portion of the identifier is actually unnecessary and could be shorted to just such as “New_York” – or put another way the city name itself is an unambiguous identifier (as the database is currently defined).  Do we wish to use the city name only as the identifier?

 

3) Should we support any level of “timezone inheritance”.  Potentially options:

    a) no inheritance: all abstime objects must declare a timezone facet (and what happens if not defined?)

    b) document level inheritance: timezone facet is inherited from parent object within document (unlike anything we do today)

    c) site level inheritance: timezone facet is inherited from a property on the Lobby object

 

Brian