Re: [Ltru] RE: ISO 639-2 decision: "mis"

["Mark Davis" <[e-mail]>]

I agree strongly with Peter. Let's suppose I have a SOAP service using the
BCP 47 protocol for communicating UI language preferences. My application
doesn't support (surprise) all possible BCP 47 language subtags, and
communicates an error back to the requester indicating when I don't accept a
particular language. Such action doesn't make me uncompliant.

The key for compliance is that if I say I'm interpreting a field according
to BCP 47,

   - I do not accept anything that isn't BCP 47 compliant,
   - Anything that I accept has to be interpreted according to BCP 47
   semantics.

There are other kinds of reasonable claims that I can make. I can say, for
example, that I interpret a field according to BCP 47, but with the addition
of 4 exceptions for backwards compatibility with previous versions of my
software.

Mark

On 6/15/07, Peter Constable <[e-mail]> wrote:
>
> You appear to be assuming that for something to qualify as "an application
> of BCP 47" it must support all tags defined by BCP 47.
>
> I would certainly agree that a BCP 47-conformant parser must fully support
> the syntax that's defined, and for a BCP-47 validator to validate all and
> only tags that are defined in relation to a given snapshot of the registry.
>
> But I see no reason why a higher-level protocol cannot restrict the set of
> tags it accepts and still be considered "an application of BCP 47". I
> wouldn't be surprised if was actually common for people to create
> higher-level protocols that normatively reference BCP 47 but that define
> some subset of tags that are permitted. What is the problem in saying those
> are "applications of BCP 47"? It would certainly be getting applied in those
> cases.
>
>
> Peter
>
>
> -----Original Message-----
> From: Randy Presuhn [mailto:[e-mail]]
> Sent: Friday, June 15, 2007 11:16 AM
> To: LTRU Working Group
> Subject: Re: [Ltru] RE: ISO 639-2 decision: "mis"
>
> Hi -
>
> As a technical contributor...
>
> > From: "Peter Constable" <[e-mail]>
> > To: "LTRU Working Group" <[e-mail]>
> > Sent: Friday, June 15, 2007 11:02 AM
> > Subject: RE: [Ltru] RE: ISO 639-2 decision: "mis"
> ...
> > MARC is an application context -- an application of language tags, not
> an
> > application in the sense of a piece software. In that application
> context, only
> > IDs from 639-2 can be used: that is a constraint on the range of
> language tags
> > supported in that application context.
>
> As such, I'd argue that these constraints mean it isn't an application of
> BCP 47.
>
> > Whether a user or a software process is assigning tags, the options are
> not
> > necessarily binary, as you suggest: in that application context it is
> entirely possible
> > for the language of given content to be known but not to be within the
> restricted set
> > permitted by MARC.
>
> These things might be called tags, and they might even be a subset of the
> set of strings that are well-formed, valid BCP 47 tags, but this is *NOT*
> an
> application of BCP 47, and, consequently, is not our problem.
>
> Randy
>
>
>
> _______________________________________________
> Ltru mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/ltru
>
>
> _______________________________________________
> Ltru mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/ltru
>



-- 
Mark

_______________________________________________
Ltru mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/ltru