Re: Possible BofF question -- I18n (was: Re: Possible OBF question -- I18n)

[John C Klensin <[e-mail]>]

--On Thursday, May 31, 2018 08:27 +0200 Patrik Fältström
<[e-mail]> wrote:

> On 31 May 2018, at 4:56, John C Klensin wrote:
> 
>> So my alternate hypothesis and reason for floating the BOF
>> idea starts by assuming that, whether people can find some
>> small amount of support or are dumb and/or committed enough
>> to do the work anyway, the needed documents can and will be
>> written and hence that the problem is getting adequate review
>> to convince the IESG and the IETF community that those
>> documents have been sufficiently checked and vetted to make
>> publication --including as standards track when appropriate
>> -- plausible and a reasonable expectation.
> 
> The serious question here is if IETF do have enough competence
> in I18N space or if IETF should drop that ball and give to
> some other SDO.

I think that leads to two other questions that I've at least
been hinting at.  

(1) If that is the right answer, how do we make it?  Personally,
I don't want to hear it via an announcement that the IEASBG
decided, entered into some sort of negotiation, and is telling
the community what they have done.

(2) Did you have a candidate SDO in mind?  Without any
disrespect to the SDOs or people involved with them, I think
we've seen ample evidence that, e.g., an SDO that is primarily
concerned with character coding and closely-related issues is
unlikely to be willing and able to understand complex issues
with the Internet and Internet protocols and prioritize them in
a way we would find acceptable.  One clear example of that
involves whether identifiers can actually be
language-insensitive.   Perhaps the answer is actually "no" but,
if so, we have made a number of bad protocol decisions that will
need to be untangled if things are ever going to work well and
I'd be surprised if it would be wise, or the IETF would be
willing, to hand those off.  Consider the DNS, identifiers for
messaging protocols, X.509 certificate identifiers, email
addresses, and so on.

--On Thursday, May 31, 2018 09:37 -0400 John R Levine
<[e-mail]> wrote:
 
> I18N needs an unusual combination of computer and linguistic
> skills.  I would be surprised if there were a lot of any
> people anywhere who have them both.

I think the list may be longer than that, with both human and
artificial interaction design (including UI, but not limited to
UI design) high on the list.   Given industry trends in user
input, especially on mobile platforms, probably speech
recognition too.   Probably you include those in "computer
skills", but I think spelling it out a bit more makes the
problem more clear.

Maybe it is a strange idea, but, if someone is going to screw up
the Internet in this area, I'd rather have it be the IETF than
some body whose major concerns in practice are not a better
Internet (whether those concerns are a particular technical
focus or economics).

I could see a "joint development" collaboration between the IETF
and some other SDO(s) working together and cross-checking each
other's work --not merely cross-area, but cross-SDO-- but that
IETF has never been able to pull one of those off before and,
AFAIK, has never expressed much interest in doing so.  Even if
we went that route, we might need a different review mechanism
than throwing a spec into the air with an IETF LC and seeing
what came down.

    john