fate of IRI working group in IETF

[Larry Masinter]

During the last week and this, there were a number of discussions about the various efforts around the IRI specification and conflicts (or potential conflicts, or relationships) between the IETF RFCs 3986, 3987, 4395, and IETF IRI wg drafts for 3987bis, comparison, registration, bidi, the W3C HTML working group documents which currently refer to them, http://url.spec.whatwg.org in WHATWG, and the planned W3C WebApps working group URL spec which is planned to produce a stable copy of the WHATWG spec.dd

I thought I'd would state my personal opinions, for the record, not as an official statement from anyone (Adobe, W3C, IETF, ...), but just in case I might have been misquoted (or misspoke):e

Background:
* I have been disappointed by the lack of participation, energy, document reviews in the IETF IRI working group; a working group which doesn't have significant active participation and document review should be shut (rather than trying to maintain the illusion of progress)
* there are significant parts of the Internet infrastructure (including most of HTTP protocol-based servers proxies gateways) which use RFC 3986 as is, and do not want or need IRIs or a broader definition
* most of the implementors of browsers are interested in following and giving feedback on http://url.spec.whatwg.org .


Given this background:
* I support closing IETF IRI working group and discontinuing work on its documents. If for some reason it remains open, or others wish to continue work in IETF on these specs as independent submissions, please remove me as editor.

* I support encouraging Anne van Kesteren to continue working on http://url.spec.whatwg.org

* I support the plan that the WebApps working group will produce a derivative work which is a stable edition suitable for reference by a stable document such as the W3C HTML standard (having stable documentation for critical system interfaces is necessary for many procurement contracts).

* I encourage Anne and others to continue to work on the URL spec, testing and test cases, and considering including more of the useful work done in the current IRI working group documents.

* If and when a stable URL document is produced (via WebApps WG) I will encourage an IETF RFC to "close off" the IRI standards track by obsoleting RFC 3987 with a reference to the W3C stable URL specification (if feasibile).

* While liaison between IETF and W3C is good, I wish others had taken a more proactive approach to resolving the apparent organizational conflicts, including the WHATWG perspective. There are a few other issues caught between the organizations (the "willful violations"), and I hope that a calmer more principled approach can be taken.

Questions? Comments?

Larry
--
http://larry.masinter.net


> -----Original Message-----
> From: "Martin J. Dürst" [mailto:***@it.aoyama.ac.jp]
> Sent: Thursday, November 08, 2012 5:40 AM
> To: Peter Saint-Andre
> Cc: public-***@w3.org
> Subject: Meeting materials (was: Re: updated IRI agenda)
>
> On 2012/11/07 2:50, Peter Saint-Andre wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Based on discussion with our Area Director, I have updated the agenda
> > so that we cover the "URL" topic and future path topic first. It would
> > not surprise me if those discussions run over the allotted time, so be
> > prepared for the possibility that we will not get to some of the later
> > agenda items.
>
> That's what happened. It's a bit sad, because although not overly
> impressive, there was actually some progress.
>
> Also, the agenda did not include a link to the meeting materials at
> https://datatracker.ietf.org/meeting/85/materials.html#wg-iri
> (I'm not blaming Peter, because there are way more links in the Agenda
> than what I'd be able to collect.)
>
> Of the four slide links, I think the Introduction slides and the "Recent
> Bulk Registration Experience" slides
> (http://www.ietf.org/proceedings/85/slides/slides-85-iri-8.pdf) were shown.
>
> I hope that everybody interested can have a quick look at the slides on
> the main draft (draft-ietf-iri-3987bis,
> http://www.ietf.org/proceedings/85/slides/slides-85-iri-6.pdf, 8 pages)
> and the slides on the bidi draft
> (http://www.ietf.org/proceedings/85/slides/slides-85-iri-7.pdf).
>
> Regards, Martin.
>
> > Agenda bashing on the list or in real time this evening
> > is of course welcome.
> >
> > http://www.ietf.org/proceedings/85/agenda/agenda-85-iri
> >
> > Peter
> >
> > - --
> > Peter Saint-Andre
> > https://stpeter.im/
> >
> > -----BEGIN PGP SIGNATURE-----
> > Version: GnuPG/MacGPG2 v2.0.18 (Darwin)
> > Comment: Using GnuPG with Mozilla - http://www.enigmail.net/
> >
> >
> iEYEARECAAYFAlCZTeIACgkQNL8k5A2w/vy4awCfXHp5HS+fbKFdPSXRW1qhnksR
> > 4rgAn2n1om/th6ouVvZLx+x0noeHhAx+
> > =huJe
> > ----

[SM]

Hi Larry,
At 17:49 08-11-2012, Larry Masinter wrote:
>Background:
>* I have been disappointed by the lack of participation, energy,
>document reviews in the IETF IRI working group; a working group
>which doesn't have significant active participation and document
>review should be shut (rather than trying to maintain the illusion
>of progress)

Working Groups discussing about internationalization tend to suffer
from lack of participation. I listened to the arguments about the
importance of doing the work. It doesn't change the fact that there
isn't any energy to do the work.

Regards,
-sm

[Jiankang YAO]

----- Original Message -----
From: "SM" <***@resistor.net>
To: "Larry Masinter" <***@adobe.com>
Cc: <public-***@w3.org>
Sent: Tuesday, November 13, 2012 6:03 PM
Subject: Re: fate of IRI working group in IETF


> Hi Larry,
> At 17:49 08-11-2012, Larry Masinter wrote:
>>Background:
>>* I have been disappointed by the lack of participation, energy,
>>document reviews in the IETF IRI working group; a working group
>>which doesn't have significant active participation and document
>>review should be shut (rather than trying to maintain the illusion
>>of progress)
>
> Working Groups discussing about internationalization tend to suffer
> from lack of participation. I listened to the arguments about the
> importance of doing the work. It doesn't change the fact that there
> isn't any energy to do the work.
>
>

it is an important work, but why do few people paritcipate in this WG?

is it due to that the importance of this work is not recognized by every involved person?

[SM]

Hi Jiankang,
At 18:35 03-01-2013, Jiankang YAO wrote:
>it is an important work, but why do few people paritcipate in this WG?
>
>is it due to that the importance of this work is not recognized by
>every involved person?

It is difficult to find people with the relevant expertise. The
people can be busy. Saying that the work is important does not change that.

Regards,
-sm

[Martin J. Dürst]

On 2013/01/04 15:17, SM wrote:
> Hi Jiankang,
> At 18:35 03-01-2013, Jiankang YAO wrote:
>> it is an important work, but why do few people paritcipate in this WG?
>>
>> is it due to that the importance of this work is not recognized by
>> every involved person?
>
> It is difficult to find people with the relevant expertise. The people
> can be busy. Saying that the work is important does not change that.

Yes. And because it's important for everybody, everybody can think that
somebody else will do it, so they don't have to do it. Also, it's
important to everybody, but except for boundary cases (error processing,
i18n, bidi) it "just works" and isn't a "major pain point".

Regards, Martin.

[John C Klensin]

--On Thursday, January 03, 2013 22:17 -0800 SM <***@resistor.net>
wrote:

> Hi Jiankang,
> At 18:35 03-01-2013, Jiankang YAO wrote:
>> it is an important work, but why do few people paritcipate
>> in this WG?
>>
>> is it due to that the importance of this work is not
>> recognized by every involved person?
>
> It is difficult to find people with the relevant expertise.
> The people can be busy. Saying that the work is important
> does not change that.

There are at least two other hypotheses:

(1) Even though there is general agreement that
internationalization is very important, reasonable people can
disagree about what should be internationalized (or localized)
and how. Several regularly-repeated discussions are of pieces
of that issue. For example:

* What constitutes a "protocol identifier" that should
not be internationalized.

* Although it is rarely discussed, it is often been
observed that, when "meaning" is not important, basic
Latin characters are understood by most of the world's
population and can be rendered by most of the world's
devices. They are, so far, required by most things that
are clearly protocol identifiers (such as URI scheme
names) so that inability to render them is a problem
regardless of what is done about i18n globally. From
that perspective, allowing other character sets globally
tends to fractionalize the Internet, not unify and
internationalize it.

* Some activities are inherently local and a matter of
localization, not subjects for i18n. For example
keyboard mappings are inherently local -- no one serious
has proposed an "internationalized keyboard" with enough
keys and shifts to be able to represent all of Unicode
(or even all abstract letters and digits in Unicode)
without escape conventions.

* There is often a useful distinction between a thing,
the name by which the thing is called, and mechanisms
that may lead to the thing. The distinction recently
drawn in the "new URL standard" thread between URL
processing and strings that may lead to URLs is a useful
part of that discussion, but so are the "to map or not"
discussions about strings that could be construed as
IDNs and the issues surrounding whether end users really
use domain names or are (or should be) using search
engines and other "above DNS" or "non-DNS" approaches.

Those are just examples and each involves tradeoffs but, if
someone examines even one of them and concludes that IRIs are
the wrong solution to the problem (or a solution to the wrong
problem), then they can conclude that IRIs are not particularly
important even if i18n is.

(2) As soon as the IRI WG started down the path of saying "these
are protocol identifiers, mostly important for protocols that
have not yet been defined in URL terms" (note that, while I hope
that is a reasonable characterization of a position, I am not
claiming that it is a consensus one or that it represents the
consensus of the active participants in the WG or of the
community), then the importance of IRIs becomes related to
guesses about protocols not yet designed, not the Internet (or,
especially the Web and URLs) as we know it today.

Those three reasons -- the two above and the issues of time,
personal or business priority among the experts, and "pain
points" that SM and Martin identifies-- are largely independent
of each other but probably have an additive effect in reducing
the number of people who are enthused about IRIs and willing to
spend major energy on them.

best,
john

[Larry Masinter]

John and SM:

The question and your interests presume the lack of interest is in the topic. But this
is not the case. Rather, There is a lack of interest in working within the IETF
process.

There ARE people willing to work, in a standards working group, on IRIs and related items.
We have the WHATWG spec:

http://url.spec.whatwg.org/

If you search http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/ for "URL standard"
And you'll find hundreds of emails about the URL spec.

We have
http://dvcs.w3.org/hg/url/raw-file/default/Overview.html

and searching the mail archive:
http://www.w3.org/Search/Mail/Public/search? hdr-1-name=subject&hdr-1-query=url&type-index=public-webapps
again with hundreds of emails about the URL spec there.

And we have
http://www.w3.org/html/wg/drafts/html/master/infrastructure.html#urls
and related bug reports and discussion about that.

I think some people might have a plan for converging these, but I can't find any written description
of the plan.

In summary: your reasoning is interesting, but does not apply to the
current situation.

Larry
--
http://larry.masinter.net


> -----Original Message-----
> From: John C Klensin [mailto:john-***@jck.com]
> Sent: Sunday, January 06, 2013 6:25 AM
> To: SM; Jiankang YAO
> Cc: public-***@w3.org
> Subject: Re: fate of IRI working group in IETF
>
>
>
> --On Thursday, January 03, 2013 22:17 -0800 SM <***@resistor.net>
> wrote:
>
> > Hi Jiankang,
> > At 18:35 03-01-2013, Jiankang YAO wrote:
> >> it is an important work, but why do few people paritcipate
> >> in this WG?
> >>
> >> is it due to that the importance of this work is not
> >> recognized by every involved person?
> >
> > It is difficult to find people with the relevant expertise.
> > The people can be busy. Saying that the work is important
> > does not change that.
>
> There are at least two other hypotheses:
>
> (1) Even though there is general agreement that
> internationalization is very important, reasonable people can
> disagree about what should be internationalized (or localized)
> and how. Several regularly-repeated discussions are of pieces
> of that issue. For example:
>
> * What constitutes a "protocol identifier" that should
> not be internationalized.
>
> * Although it is rarely discussed, it is often been
> observed that, when "meaning" is not important, basic
> Latin characters are understood by most of the world's
> population and can be rendered by most of the world's
> devices. They are, so far, required by most things that
> are clearly protocol identifiers (such as URI scheme
> names) so that inability to render them is a problem
> regardless of what is done about i18n globally. From
> that perspective, allowing other character sets globally
> tends to fractionalize the Internet, not unify and
> internationalize it.
>
> * Some activities are inherently local and a matter of
> localization, not subjects for i18n. For example
> keyboard mappings are inherently local -- no one serious
> has proposed an "internationalized keyboard" with enough
> keys and shifts to be able to represent all of Unicode
> (or even all abstract letters and digits in Unicode)
> without escape conventions.
>
> * There is often a useful distinction between a thing,
> the name by which the thing is called, and mechanisms
> that may lead to the thing. The distinction recently
> drawn in the "new URL standard" thread between URL
> processing and strings that may lead to URLs is a useful
> part of that discussion, but so are the "to map or not"
> discussions about strings that could be construed as
> IDNs and the issues surrounding whether end users really
> use domain names or are (or should be) using search
> engines and other "above DNS" or "non-DNS" approaches.
>
> Those are just examples and each involves tradeoffs but, if
> someone examines even one of them and concludes that IRIs are
> the wrong solution to the problem (or a solution to the wrong
> problem), then they can conclude that IRIs are not particularly
> important even if i18n is.
>
> (2) As soon as the IRI WG started down the path of saying "these
> are protocol identifiers, mostly important for protocols that
> have not yet been defined in URL terms" (note that, while I hope
> that is a reasonable characterization of a position, I am not
> claiming that it is a consensus one or that it represents the
> consensus of the active participants in the WG or of the
> community), then the importance of IRIs becomes related to
> guesses about protocols not yet designed, not the Internet (or,
> especially the Web and URLs) as we know it today.
>
> Those three reasons -- the two above and the issues of time,
> personal or business priority among the experts, and "pain
> points" that SM and Martin identifies-- are largely independent
> of each other but probably have an additive effect in reducing
> the number of people who are enthused about IRIs and willing to
> spend major energy on them.
>
> best,
> john
>

[JFC Morfin]

At 19:44 06/01/2013, Larry Masinter wrote:
>The question and your interests presume the lack of interest is in
>the topic. But this is not the case. Rather, There is a lack of
>interest in working within the IETF process.

This is a very interesting statement when I confront it to my
personal field current experience.

The IETF' switch from the better internet to the more market
palatable internet (Aug 29th) might actually correspond to this. Like
a change wehad before from ITU to IETF. This time it is from
internationalization to multilingualization (as a general metaphore
in many disciplines). It might not mean many new SDOs but a possible
switch from internationalizing (Unicode, IETF, etc.) to
multilingualizing entities, actually from localization to subsidiarity.

I personnally locate the change at RFC 5895. The innetwork is stable,
reliable, etc. the outnetwork is adaptative, at least on a community
basis. Languages are good examples of existing communities.
Technologies may also be (ipads, linux, etc). Architures may also be.
I experiment the concept with the small IUse community slow emergence
(Intelligent Use system extensions).

jfc