Re: [Ltru] Geocoordinates

It is not necessarily non-conformant. I could define a field as follows:

   - if it has 4 characters and no hyphen, it means X
   - otherwise it is interpreted as a BCP 47 language tag.

That is perfectly conformant usage of BCP 47, and would work with the
current spec. The disadvantage is that if BCP47 ever uses 4-letter base
language codes, then you can't represent the full range of BCP47 at that
point. So you definitely take a risk in using such a definition in terms of
future compatibility, but such a definition is not non-conformant.

Mark

On Wed, Mar 11, 2009 at 10:31, Randy Presuhn
<randy_presuhn@mindspring.com>wrote:

> Hi -
>
> > From: "Peter Constable" <petercon@microsoft.com>
> > To: "John Cowan" <cowan@ccil.org>; "Phillips, Addison" <
> addison@amazon.com>
> > Cc: "LTRU Working Group" <ltru@ietf.org>
> > Sent: Tuesday, March 10, 2009 8:36 PM
> > Subject: Re: [Ltru] Geocoordinates
> >
> > From: ltru-bounces@ietf.org [mailto:ltru-bounces@ietf.org] On Behalf Of
> John Cowan
> >
> > > So what's left is:
> > >
> > > (a) codes for language collections that aren't in 639-5, which could
> > > be added to BCP 47 as 4-letter primary language subtags (that size is
> > > currently reserved);
> >
> > I'd be resistant to 4-letter primary language subtags.
> > I have made derivative use of BCP 47 tag syntax
> > with the modification that the only obligatory subtag
> > is the script subtag.
> > Introducing different 4-letter subtags within BCP 47 syntax
> > would create conflicts with this or other similar derivative
> > uses of our language tag syntax.
>
> As a technical contributor...
>
> I find this line of argument unpersuasive.  We should only
> concern ourselves with compatibility with conformant
> implementations.  If someone implements something
> "derived from" the spec which does not follow it to the
> letter, they're on their own.
>
> Randy
>
> _______________________________________________
> Ltru mailing list
> Ltru@ietf.org
> https://www.ietf.org/mailman/listinfo/ltru
>