Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

Phillips, Addison wrote:

> But ‘northern’ wouldn’t meaning anything different. It is one thing to 
> allow a generic subtag to mutate meaning differently in different 
> contexts. It is another thing to have a totally different meaning 
> accrete to a subtag. If ‘1901’ meant both the German orthography and, 
> say, some Uzbek transliteration scheme, you couldn’t tell which was 
> meant.

I agree.  I can see 'pinyin' being used for slightly different pinyins, 
but not when the two meanings are altogether different.

> Type: variant
> Subtag: 1901
> Description: German orthography of 1901
> Description: Uzbek transliteration of 1901
> Prefix: de
> Prefix: uz
> Comments: Good luck figuring this out.

I agree, this would be bad.  But even worse would be:

Type: variant
Subtag: 1901
Description: German orthography of 1901
Prefix: de
Comments: Don't mix this up with the other 1901.
%%

Type: variant

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

> 
> > I also note that no truly generic variants (such as northern)
> have
> > been accepted. Both fonipa and fonupa refer to something concrete
> that
> > could be applied to (m)any languages.
> 
> We've traditionally used "generic" to mean "applicable to all
> (languages)" and not "abstract."

"Abstract" would be a better word in this context. So: we have not registered any abstract generic variants 

I would say, though, that generic variants are not *necessarily* applicable to 'all' languages. It is that
the list of languages is non-specific. One could possibly create an enumeration, but that list might be
quite large and it isn't known in advance.

Addison

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

Hi -

As a technical contributor...

> From: "Mark Davis" <mark <at> macchiato.com>
> To: "John Cowan" <cowan <at> ccil.org>
> Cc: "LTRU Working Group" <ltru <at> ietf.org>
> Sent: Tuesday, September 30, 2008 11:25 PM
> Subject: Re: [Ltru] Uniqueness of variant subtags (Was: BCP47 Appeals process
>
> As am I.
> 
> I'm not opposed to
> 
> de-1901
> uz-1901
> and so on. These are meaningful and easy for people to understand.
> 
> I think that can be encapsulated as:
> 
> Type: variant
> Subtag: 2003
> Description: Denotes a change in orthography, transliteration, or other
> usage associated with the publication of a significant standard or other
> specification in 2003
> Prefix: de
> Prefix: uz
> Prefix: be-ungegn
> 
> We can then add prefixes that are known to have such cases over time.

I'm worried about "fuzziness" creeping in  in cases like this.
A registration request might provide a very precise definition,
identifying a particular national or international standard
for a particular prefix.  The variant record, lacking this
information, might consequently result in the subtag being
interpreted much more broadly in the case of that particular
prefix.  (That's the crux of my concern with respect to
"pinyin" on the ietf-languages <at> iana.org list, especially
when understood for extended pinyin-like schemes, since
its entirely conceivable that more than one might exist
for languages for which pinyin wasn't originally designed.)

The net result, then, is that yet another variant becomes
necessary for the strictly defined variant, bringing with
it potential user confusion.  (E.g., if it's really
Hanyu Pinyin, tag it zh-cmn-Latn-pinyin-hpinyin?)

Randy

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

Randy Presuhn 2008-10-01 08.52:

> As a technical contributor...
>> From: "Mark Davis" <mark <at> macchiato.com>

>> I'm not opposed to
>> 
>> de-1901
>> uz-1901
>> and so on. These are meaningful and easy for people to understand.
>> 
>> I think that can be encapsulated as:
>> 
>> Type: variant
>> Subtag: 2003
>> Description: Denotes a change in orthography, transliteration, or other
>> usage associated with the publication of a significant standard or other
>> specification in 2003
>> Prefix: de
>> Prefix: uz
>> Prefix: be-ungegn
>> 
>> We can then add prefixes that are known to have such cases over time.
> 
> I'm worried about "fuzziness" creeping in  in cases like this.
> A registration request might provide a very precise definition,
> identifying a particular national or international standard
> for a particular prefix.

"First to mill" doesn't seem the right principle when it comes to 
the right to "occupy" number "2003" - referring to year 2003 AD. 
Therefore, I agree with Mark that this makes sense.

Defining e.g. "2003" the way Mark suggests here, also creates a 
"registration pattern": It shows a way to do it.
--

-- 
leif halvard silli

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

At 05:39 08/10/01, Stephane Bortzmeyer wrote:
>On Tue, Sep 30, 2008 at 08:40:36AM -0700,
> Phillips, Addison <addison <at> amazon.com> wrote 
> a message of 41 lines which said:
>
>> Since there is no consensus on my proposed change, unless directed
>> by the chairs otherwise, I will removed my proposed text from the
>> editor's copy before incorporating any other changes.
>
>True, there is no consensus yet but the issue is a real one, no? RFC
>4646 does not provide clear guidance (although ietf-languages'
>practice has always been to refuse identical variant
>subtags). Shouldn't we delay the I-D a bit for that?

I hope we can restrict this to the a very short time.
Everybody, please comment quickly and express clearly
what you can absolutely not live with vs. just desirable
stuff, so that we can converge quickly.

Regards,    Martin.

#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst <at> it.aoyama.ac.jp     

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

I'm not really sure what's wrong, in terms of "figuring it out" with:

Type: variant
Subtag: 1901
Description: German orthography of 1901
Description: Uzbek transliteration of 1901
Prefix: de
Prefix: uz

(it should be clear that the 'de' is German and the 'uz' is Uzbek).
If this isn't clear, then what about the following:

Type: variant
Subtag: 1901
Description: German orthography of 1901
Description: Uzbek transliteration of 1901
Prefix: de
Prefix: uz
Comment: with Prefix de, German orthography of 1901
Comment: with Prefix uz, Uzbek transliteration of 1901

(but it seems kind of silly, doesn't it).

I'm not saying we should do this a lot, but I don't see any
really fundamental problem with it in terms of figuring out
what things mean, which is the most important point.

Regards,    Martin.

At 05:43 08/10/01, Randy Presuhn wrote:
>Hi -
>
>As a technical contributor...
>
>> From: "Phillips, Addison" <addison <at> amazon.com>
>> To: "Stephane Bortzmeyer" <bortzmeyer <at> nic.fr>
>> Cc: "LTRU Working Group" <ltru <at> ietf.org>; "Ewell" <doug <at> ewellic.org>; 
><Doug <at> core3.amsl.com>
>> Sent: Tuesday, September 30, 2008 8:39 AM
>> Subject: Re: [Ltru] Uniqueness of variant subtags (Was: BCP47 Appeals process
>>
>> Let me retry expressing this. Consider this record:
>> 
>> Type: variant
>> Subtag: 1901
>> Description: German orthography of 1901
>> Description: Uzbek transliteration of 1901
>> Prefix: de
>> Prefix: uz
>> Comments: Good luck figuring this out.
>
>This is precisely how the ietf-languages <at> iana.org list appears
>to proposing that "pinyin" be handled.
>
>Consider, on the other hand...
>
>Type: variant
>Subtag: 1901
>Description: German orthography of 1901
>Prefix: de
>Comments: Should be clear.
>
>Type: variant
>Subtag: 1901
>Description: Uzbek transliteration of 1901
>Prefix: uz
>Comments: Should be clear.
> 
>We should be clear about whether we want to permit this or not.
>I think there is no problem when the prefix is used.  The problem
>is figuring out what it means when one of these variants is used
>with any other prefix.  (It's also a problem if someone has used the
>subtag as a database key.)
>
>Randy
>
>_______________________________________________
>Ltru mailing list
>Ltru <at> ietf.org
>https://www.ietf.org/mailman/listinfo/ltru

#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst <at> it.aoyama.ac.jp     

Re: Uniqueness of variant subtags (Was: BCP47 Appeals process

At 06:00 08/10/01, Stephane Bortzmeyer wrote:

>Yes. It seems most developers assumed that identical variant subtags
>were forbidden and so allowing your solution would break
>compatibility.
>
>Also, for all other types, duplication of a subtag is forbidden. It
>may be difficult to explain why variants are different.

 From an implementation point of view, it may be very difficult to
explain. From a user point, it's rather easy to explain; variants
are secondary information, whereas others (incl. regions, scipts)
have a primary meaning, even if they are used for secondary
distinctions in language tags.

Regards,   Martin.

#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst <at> it.aoyama.ac.jp     

Re: Uniqueness of variant subtags

Many wrote:

>>> I am wholly and utterly opposed to allowing two records of the same 
>>> type with the same subtag.
>>
>> And so am I.
>
> As am I.

Me too.  Could not live with this.

Mark added:

> I'm not opposed to
>
> de-1901
> uz-1901
> and so on. These are meaningful and easy for people to understand.

As a matter of subtag assignment policy, which is really a question for 
ietf-languages, I think we are seriously overestimating both the need to 
tag variations identified by a year and the likelihood that users will 
connect these year numbers to the variations.

German was a notable special case: it had a well-known, widely 
publicized, and much-debated orthographic revision for which we ended up 
creating the tag "de-1996" (with complement "de-1901" for the old 
orthography) back in the RFC 3066 whole-tag days.  This was largely 
because we wanted to avoid something like "de-revised" knowing that it 
might be revised again some day.  However, even now we talk of "de-1901" 
and "de-1996" as making distinctions that usually don't need to be made; 
see, for example, Section 4.1 of both RFC 4646 and draft-4646bis-17.

Since assigning "1994" to a Resian orthography invented by Han 
Steenwijk, which might well have been named for him instead, 
ietf-languages has been more and more willing to use year numbers to 
identify language variations, even those where the variation is not 
commonly associated with the year, or where regular revisions 
identifiable by year are not expected.  We are about to register 
'1959acad' for Belarusian even though we distinguish no other revisions 
of Academy Belarusian, and we were considering 'hpin1958' for Hanyu 
Pinyin even though the distinction we wanted to draw was between Hanyu 
and other Pinyins, not between 1958 and any other revision.

I'm not saying year numbers don't ever have their place, but it worries 
me when we expect them to be so common that we are thinking about 
compromising an important property of the Registry, uniqueness of subtag 
values within type, in anticipation that there will be tons of them.

And no, I don't consider this inconsistent with my position that 'coruc' 
and 'corur' and so forth are acceptable subtags.  These may be cryptic, 
but they are not arbitrary.

--
Doug Ewell  *  Thornton, Colorado, USA  *  RFC 4645  *  UTN #14
http://www.ewellic.org
http://www1.ietf.org/html.charters/ltru-charter.html
http://www.alvestrand.no/mailman/listinfo/ietf-languages  ˆ

_______________________________________________
Ltru mailing list
Ltru <at> ietf.org
https://www.ietf.org/mailman/listinfo/ltru