Re: [Ltru] Well-formed vs. regular (Was: I-D ACTION:draft-ietf-ltru-4646bis-01.txt

["Mark Davis" <mark.davis at icu-project.org>]

I agree that the language needs a bit of cleanup, but not quite what you suggestion.

Well-formed means that it conforms to the ABNF for Language-Tag AND it contains no two identical singleton extension tags.

We're using 'regular' to mean conforms to the ABNF for langtag or privateuse, and 'irregular' to mean the reverse: doesn't conform to the ABNF for langtag or privateuse. (Thus corresponding to the language tags that must be syntactically grandfathered.)

Mark

On 12/14/06, Stephane Bortzmeyer < bortzmeyer at nic.fr> wrote:

On Thu, Dec 07, 2006 at 03:50:02PM -0500,
Internet-Drafts at ietf.org < Internet-Drafts at ietf.org> wrote
a message of 94 lines which said:

>       Title           : Tags for Identifying Languages
>       Author(s)       : A. Phillips, M. Davis
>       Filename        : draft-ietf-ltru-4646bis-01.txt

The draft seems inconsistent in its use of "well-formed" and
"regular". We find:

>   grandfathered = langtag   ; well-formed grandfathered tags
>                 / irregular ; tags that are not well-formed

(I would say that irregular is the opposite of regular, not of
well-formed.)

>  Those tags that would not be well-formed according to the ABNF in
>  this document or that contain subtags that do not individually
>  appear in the registry are maintained in the registry in records of
>  the "grandfathered" type.

(These tags are legal according to the ABNF, both of RFC 4646 and of
the new draft.)

>   Some grandfathered tags are "well-formed" in that they match the
>   'langtag' production in Figure 1.

And, a few lines later:

>   An implementation that claims to check for well-formed language tags
>   MUST:
>
>   o  Check that the tag and all of its subtags, including extension and
>      private use subtags, conform to the ABNF or that the tag is on the
>      list of grandfathered tags.

(The ABNF is sufficient since the "irregular" production is in it.)

Shouldn't we stick to a consistent vocabulary? I suggest to clearly
separate "regular" (complies with the "langtag" production of the
ABNF) and its superset "well-formed" (complies with the "Language-Tag" production,
plus the restriction on repeated extensions).

With this vocabulary:

fr-Arab-AQ is well-formed (and regular)

i-klingon is well-formed (and irregular)

i-toto is not well-formed

art-lojban is well-formed (and regular)

uk-Cyrl-U is not well-formed

_______________________________________________
Ltru mailing list
Ltru at ietf.org
https://www1.ietf.org/mailman/listinfo/ltru

_______________________________________________
Ltru mailing list
Ltru at ietf.org
https://www1.ietf.org/mailman/listinfo/ltru