At 12:19 23/11/2005, Stephane Bortzmeyer wrote:
On Mon, Nov 21, 2005 at 09:54:27AM +0100, Julien.Maisonneuve@xxxxxxxxxxx <Julien.Maisonneuve@xxxxxxxxxxx> wrote a message of 25 lines which said: > The IETF is probably the ONLY meaningful organisation in the world > that insists on using ascii-only specifications. Any rationalization > of that practise should try to explain why we are so exceptional
So, saying "Everybody do it that way, we should, too" is ignoring many IETF idiosyncrasies.
Full agreement.
3) IETF attendance is quite open (here, IETF is really exceptional). So, its members are a wide and diverse community and it is difficult to find a format which is both better than ASCII and widely available.
This is here where however we have a real practical problem. Only standards and procedures written in English can be authoritative and Drafts needs to be designed using the ASCII charsets. This is acceptable when only English speaking engineers and reading users are concerned. This represents less than 1/4 of the world population. As long as the IETF addressed pure Internet technology, it was unlikely this would be a problem - and after all an ASCII technology can legitimately only documented in ASCII. This is not anymore the case with the increase of societal aspects in the digital ecosystem convergence and with multilingual issues.
The BCP 49 update the IESG has approved, and I oppose in its current form for that reason, creates and is unable to address that difficulty. As a BCP it is supposed to document practices which - by essence - can be increasingly authoritatively multilaterally documented in the concerned languages (the translation is not from an authoritative English text to another language, but from another language authoritative text to ASCII English). RFC 3066 bis engages the IETF in having to be unilaterally authoritative in areas where it has no internal expertise (and therefore no cross-expertise), no adequate tools and procedures and no proper resources (IANA is not multilingual - or even internationalized).
This is why after more drastic initial suggestions, I found the sole (IMHO) possible consensual technical proposition: it is to simply include in RFC 3066 bis a hook (two lines of additional texte) to the RFC 4151 format (URI-tags). This way, ASCII (and proposed RFC 3066 constrained format) stays authoritative by default, but an external tagging system or procedure (whatever the script) can be authoritatively referred to in an RFC, an XML page or through a library entry, in using ASCII characters. Obviously, upgrading RFC 4151 towards full IRI-tags support would be better. This approach does not address everything, but it addresses most (including protection of the users against simple racial/cultural/religious profiling meta-spam [searches in Google through language tags the author might ignore (RFC 3066 bis Security considerations) em] and language identification encryption).
Otherwise, after being quite open, as you mention it, the IETF will become exceptional in our world of increasingly protected and respected cultural diversity, as being culturally/societally exclusive and the authoritative reference (language IANA registry) of that exclusion. I doubt this is the intent, however this is/would be the practice.
jfc _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf