Re: [Ltru] RFC 4646 production "grandfathered" considered harmful

[Martin Duerst <duerst at it.aoyama.ac.jp>]

Like John, I favor choice 2.    Regards,   Martin.

At 15:19 06/09/17, John Cowan wrote:
>Section 2.2.9 of RFC 4646 says:
>
>   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.
>
>   o  Check that singleton subtags that identify extensions do not
>      repeat.  For example, the tag "en-a-xx-b-yy-a-zz" is not well-
>      formed.
>
>(I have emphasized the word OR in the first bullet point.)
>
>Unfortunately, this wording allows too much.  For example, the invalid tag
>"ra-bb-it" matches the "grandfathered" rule in the ABNF.  Therefore it
>winds up being well-formed even though it cannot be analyzed as a sequence
>of subtags and is not on the grandfathered list either.
>
>To avoid this, we can take one of two actions:
>
>1) Remove the "grandfathered" production in the ABNF altogether, and use
>the "OR" in the conformance clause to allow the irregular grandfathered
>tags (that is, those which don't match the "langtag" or "privateuse"
>productions) to be well-formed.  The danger here is that people will
>implement the ABNF only, and the grandfathered tags will become outright
>unusable rather than merely discouraged.
>
>2) Add an explicit "irregular" production in place of the "grandfathered"
>production which explicitly enumerates the 17 irregular grandfathered
>tags, thus:
>
>irregular = "en-GB-oed" / "i-ami" / "i-bnn" / "i-default"
>            / "i-enochian" / "i-hak" / "i-klingon" / "i-lux" / "i-mingo"
>            / "i-navajo" / "i-pwn" / "i-tao" / "i-tay" / "i-tsu"
>            / "sgn-BE-fr" / "sgn-BE-nl" / "sgn-CH-de"
>
>It is safe to enumerate this list explicitly, as it can neither grow
>nor shrink.  It's true that all the tags except "i-default" can become
>deprecated, but that makes no difference to well-formed processors.
>The other grandfathered tags in the registry are all well-formed already
>and do not need to be in this list.
>
>In this case the conformance clause can be simplified by omitting the
>second part of the "OR".
>
>I favor choice 2.
>
>-- 
>John Cowan                                   cowan at ccil.org
>        "You need a change: try Canada"  "You need a change: try China"
>                --fortune cookies opened by a couple that I know
>
>_______________________________________________
>Ltru mailing list
>Ltru at ietf.org
>https://www1.ietf.org/mailman/listinfo/ltru


#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst at it.aoyama.ac.jp     


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