Re: [netmod] I-D Action: draft-lange-netmod-iana-timezones-00.txt

[Andy Bierman <[e-mail]>]

On Thu, Jun 28, 2012 at 7:58 AM,
<[e-mail]>wrote:

> > -----Original Message-----
> > From: Juergen Schoenwaelder
> [mailto:[e-mail]]
> > Sent: 28 June 2012 14:12
> > To: Lange, Jeffrey K (GE Energy); Martin Bjorklund; [e-mail]
> > Subject: Re: [netmod] I-D Action:
> draft-lange-netmod-iana-timezones-00.txt
> >
> > On Thu, Jun 28, 2012 at 03:02:30PM +0200, Juergen Schoenwaelder wrote:
> >
> > > > So that also is a bit arbitrary as well.  Should I leave the enums
> in
> > alphabetical order? Or should it match what the zone.tab file does?
> > >
> > > Any order is likely irrelevant over time. See below.
> > >
> >
> > Well, I realize this statement is not really helpful / correct. You
> > can keept the enums defined in the order they are defined in the TZ
> > database but then the value statements MUST be used and the numbers
> > over time won't be assigned in a strictly increasing order. This means
> > that YANG tools can use the definition order (if they are able to
> > access it) or they might end up using the numeric value order, which
> > will over time be increasingly meaningless.
>
> Still learning here, but RFC6087 states:
>
> If extensibility of enumerated values is required, then the
> 'identityref' data type SHOULD be used instead of an enumeration or
> other built-in type.
>
> So why are enums being considered over identityrefs?
>
>
That text should be clarified.  If a centralized naming authority
is used, then enumeration is OK, and less complex than identityref.
For distributed naming authority, an identityref is needed.
(Identityref is a QName -- module-prefix:identity-name).

Also, enums are allowed to have many chars in them that
would be invalid in a YANG identifier. (e.g. '/' would make every
timezone location invalid)


>
> > /js
> >
>


Andy