IANA registration issues not covered by the "RFC 3066bis" draft

JFC (Jefsey) Morfin jefsey at jefsey.com
Sat Jan 29 04:02:42 CET 2005

Dear John,
as you know the IETF is building an agreement with ISOC to reorganize the 
Secretariat of the Internet standard process. There are cons and pros as in 
every new proposition. However, we can probably survive a difficult IETF 
Secretariat transition. Not a similar problem at the IANA. I fully 
understand that you guys were cosy in here. But there are several major 
changes, some you introduced.

- the IETF is reforming
- the ICANN will not be renewed the USG IANA contract and fights against 
the rest of the world to keep it from its own legitimacy
- you have introduced the RFC 3066bis uncomplete process and refused amend it.
- years of International unanimous Gov declarations on multilingualism are 
starting to be questionned all over the world: are the Govs not able to be 
listen by a few experts.

There is a story which says "my wife and me lived happy for 20 years, until 
we met." Your language bachelorship time are over. You chose to meet the 
world with the RFC 3066bis draft, through W3C URI/IRI definition. This does 
not mean that the situation will be worse if we organize it well. But we 
have to organize it, and for that to know exactly where we stand.

At 14:38 28/01/2005, John Cowan wrote:
>JFC (Jefsey) Morfin scripsit:
> > True. But Michael Everson is actually carrying three tasks:
> > - Reviewer on behalf or IETF as designated by one Application AD (as per
> > RFC 3066) or both? Assisted by a non-IETF list?
>Yes.  You should also add that his decisions may be appealed to the IESG
>under RFC 2026, and that no such appeal has ever been made.

Not really clear. Because this process is only defined by RFC 3066 and not 
as an IETF WG or responsibility. For example, there are two Application 
ADs. Only one is involved: which one? But the appeal would only be against 
an OK of the reviewer. Not against the OK of the IANA registrar - what if 
IANA does not remove an entry (they have often a one year backlog plus 
policy: ask Pitcairns people). The attitude of Michael is also concerning: 
when one wants to help him, he starts "trolling" you. I fully understand 
his point of view, even if I do not find it appropriate. I think all this 
should be addressed in his role's missing description, so we would not need 
to discuss it further, and have stable and appropriate procedures.

> > - Registrar of ISO 15924 - what is out of IETF scope
> > - Registrar for IANA - as per a IANA page I do not find back right now
> > (sorry, but the IANA is a mess not a menu)
>No.  The Reviewer forwards the tags he approves to IANA for registration
>there, but this process is automatic: whoever does the actual work at
>IANA does not exercise discretion.  This is in accordance with RFC 2860,
>Section 4.1.

I try to spend time sorting out the IANA system to automate copies. I will 
come back on this when I can document it better.

> > I am not referring to this. I am referring to the registration of the
> > documentation of a language tag through a single document, book or
> > authority.
>Nothing prevents the creation of authoritative mirrors of the IANA
>langtag registry, but then nothing prevents the creation of authoritative
>mirrors of the IANA protocols or services registries either.  In
>practice, non-authoritative mirrors (see /etc/protocols and /etc/services
>on your local system) do the job effectively.

I am referring to the fact that a language is recorded as per M. so and so 
and documented as per his liking. There are a few people upset by some 
current registrations (nothing personal, this is just the impossibility to 
register different tags for the same language depending of the style and 
the author[ity])

> > I am afraid you will never understand this as a tag designer while I am a
> > tag user. In missing the style and authority fields your tag does not
> > permit the real needs of the words to be documented, leading to
> > conflicts.
>Our ten years of harmonious cooperation trump your hypothetical conflicts.

RFC 3066 dates from 2000 and as explained many times now, you and the world 
have changed your coziness.

> > Your tag does not scale and is therefore inadequate to real
> > life and violating RFC 1958.
>On the contrary.  I refer you particularly to Section 3.7 of that RFC.

Certainly and this is what Harald did in 2000. But temporary solutions can 
only last a time.

Your approach was reducing but probably acceptable for the American 
academic Internet of 1981, it survived till now in not being an open IETF 
function. Your draft made it public and mandatory through W3C. Section 3.3 
is now the reference (sites may want to chose among millions of language 
flavors). I proposed again and again you could keep the 3.7 approach in 
just stating it. This is what John Klensin and others proposed in different 
ways. Addison told me "no" refusing to discuss the other points, I was 
called a "troll" and my remarks "gerrymandering" and John Klensin's 
proposition labeled as "odious".

Just something: languages belong to 6.2 billions people what averages to a 
minimum of 62.000 languages as people tend to diversify and one commonly 
accepts that a language cannot survive with less than 100.000 speakers (but 
that figure may drastically be reduced with computer assistance and in 
special cases). Also a same person can speak several languages. So I would 
think the real need is for a flexible (intuitive, i.e. no reviewer) tagging 
of a minimum around 150.000 languages, some of them with different scripts.


>Winter:  MIT,                                   John Cowan
>Keio, INRIA,                                    jcowan at reutershealth.com
>Issue lots of Drafts.                           http://www.ccil.org/~cowan
>So much more to understand!                     http://www.reutershealth.com
>Might simplicity return?                        (A "tanka", or extended haiku)

More information about the Ietf-languages mailing list