Re: [apps-discuss] timezone database maintenance

[Cyrus Daboo <[e-mail]>]

Hi Marc,

--On March 3, 2011 11:58:55 AM -0500 Marc Blanchet 
<[e-mail]> wrote:

> A) XML IANA
> - IANA has been converting its "old" registries to XML native format.
> - for new registries, XML is used as native format.
> - TZ database has its own format.
> - wonder if it would be relevant to have IANA do the registry in XML (as
> others and to be managed by the same tools) and then convert to the TZ
> database format.

CalConnect has been working on an XML representation for timezone data, but 
I think it is fair to say at this time we consider this premature. One of 
the key things CalConnect wants to do is promote use of a Timezone Service 
protocol (draft-douglass-timezone-service about to be updated) to allow for 
timely, secure, reliable update of timezone information. Right now that is 
focussed primarily on iCalendar timezone data, but we intend to allow it to 
be used by anyone that needs timezone information, including OS's (i.e. 
allow "live" updates to zoneinfo files rather than requiring OS 'patches').

What I think we need to do now is treat the Olson data format as 
"normative" and then work on defining a solid XML representation (with 
hopefully a straightforward mapping to Olson data). We can then make that 
available as an alternative format and allow tools to adopt it, and once we 
have good adoption then move to have that be the normative format. However, 
that is something that will need more time to develop than I think we have 
before we need IANA to start managing the Olson process (given what I know 
about the length of the IETF process).

> C) Reference code.
>
> I'm not sure IANA is the right place to archive software code.  why not
> have a separate web site/sourceforge/google code/... place for managing
> this?

That would probably need additional text describing how that is managed 
(who has commit rights, how would those be administered/changed etc) and 
would be a significant change to the current Olson process. It may be 
reasonable to put the code up on a source site, but if we do that, I do 
think it is important that appropriate "snapshots" be tied to the updates 
to the IANA data - so I still think IANA ought to maintain a tarball of the 
code appropriate to the actual data.

-- 
Cyrus Daboo