<inline> Tom Petch
----- Original Message ----- From: "Sam Hartman" <[EMAIL PROTECTED]> To: "David Harrington" <[EMAIL PROTECTED]> Cc: "'tom.petch'" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Thursday, February 01, 2007 7:43 PM Subject: Re: [Syslog] An early last call comment on protocol-19 > >>>>> "David" == David Harrington <[EMAIL PROTECTED]> writes: > > David> Hi WG, If ISO is a subset of what is covered by BCP047, > David> then would it be acceptable to REQUIRE the ISO subset > David> mandatory-to-implement-for-compliance for interoperability > David> purposes, and implementations MAY support other languages > David> in BCP047 with no assurance of interoperability with > David> standard-compliant implementations? > > No, I'd really need a fairly strong justification that went through > the languages you were not supporting and explained why that was > appropriate for syslog. > > BCP 47 is by definition the IETF's best current practice for language > tagging. Absent a compelling reason to do something else, you should > identify languages that way. Tom has not (so far) presented a > compelling reason. > Pity, I had hoped that David's compromise would be acceptable. RFC4646 (the current BCP0047) is a magnificent piece of work and does enable the generator of text to specify quite precisely how it should be interpreted. I love the differentiation between the dotted letter I of Azerbaijan and Turkey, in fact all the comments about Azerbaijani, Mongolian and Icelandic. What concerns me is conformance, what does it mean that a parameter MUST conform to this BCP or any other, an issue that has surfaced on this list before. If we just changed the reference so that the I-D were to read "it MUST contain a two letter language identifier as defined in BCP0047 [13}" then I have no problem but this does rather negate the intent of the BCP. The BCP defines two levels of conformance (s.2.2.9) and I suspect that even the lower level requires online access to the IANA website so what does a receiver of a syslog message do? Take it as an opaque character string? Check the ABNF? Do as RFC4646 specifies, for well-formed or validating conformance? I suggest anyone considering this question look at the current online registry as well as RFC4646, and note such comments as '"en-a-bbb-a-ccc" is invalid' whereas 'the tag "en-a-bbb-x-a-ccc"' is valid; or "sl-IT-nedis" is suitable but "it-IT-nedis" is not. It is beautiful. The issue I see is conformance. What can we expect the recipient of a syslog message to do without placing a significant burden thereon? Note too that RFC4646 defines a character string so we also have to specify an encoding thereof, another issue that has surfaced on this list before. Tom Petch _______________________________________________ Syslog mailing list Syslog@lists.ietf.org https://www1.ietf.org/mailman/listinfo/syslog
