Message-Id: <6.2.3.4.2.20050822113322.051ee100@mail.afrac.org>
Date: Mon, 22 Aug 2005 12:12:36 +0200
To: Martin Duerst <duerst@it.aoyama.ac.jp>, ltru@ietf.org
From: r&d afrac <rd@afrac.org>
In-Reply-To: <6.0.0.20.2.20050822155409.094c3430@itmail.it.aoyama.ac.jp>
References: <6.2.3.4.2.20050820023107.04315da0@mail.afrac.org>
	<6.0.0.20.2.20050822155409.094c3430@itmail.it.aoyama.ac.jp>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
Cc: 
Subject: [Ltru] Re: matching for non-RFC3066bis langtags (was: editor's copy
 of matching updated...)
Precedence: list
Sender: ltru-bounces@lists.ietf.org
Errors-To: ltru-bounces@lists.ietf.org

At 09:05 22/08/2005, Martin Duerst wrote:
>[please use new subject for new issue, thanks!]
>
>At 09:36 05/08/20, r&d afrac wrote:
> >Dear Addison,
> >I do not understand the meaning of the sentence in the 
> introduction: " Given a set of language identifiers, such as those 
> defined in [ID.ietf-ltru-registry], various mechanisms can be 
> envisioned for performing language negotiation and tag matching.".
> >
> >Either a language tags uses subtags registered in the IANA 
> registry (except "x-") and obey to the ABNF (and the "such" seems a 
> contradiction) or I do not see how you can envision mechanisms on a 
> language identification system you do not know?
>
>I think there are some very general kinds of matching that can apply
>independent of the syntax. E.g. various completely different tagging
>syntaxes may allow matching of a list of languages (e.g. the list of
>languages somebody speaks or reads) against a particular language,
>or they may have one way or another to deal with subset relationships.
>Of course it may be possible to envision language tagging syntaxes where
>even these kinds of matching don't make sense.

Dear Martin,
I am afraid there is a confusion here. What Addison talks about is of 
a set of "language identifiers" (I do not know what it is compared 
with language tags) "such as those defined in" his Draft. I 
understand that his Draft describes the structure, content, 
construction, and  semantics of language tags" to be used on the 
Internet (BCP) when a language is to be indicated. This results in a 
strict ABNF and a registry of subtags.

I do not know the extent of what you name "syntax" when referred to 
this? The only understanding I have is: a different syntax means a 
different ABNF.

But I am confused because a different syntax does not affect the 
semantic, and a different syntax should not hamper the matching as 
you note it. The only reason why such a matching would not make 
senses would be when a different language tagging would support a 
different semantic. In saying that this is possible (I suppose in 
support of the Addison's text, therefore I suppose "practically, 
technically possible under the current Draft") I feel you say (and 
Addison says in using "such") what I say while supporting a document 
which (this is the way I understand it) forbids it.

> >Or may be could you give an example of what could be another set 
> of identifier which would not be defined in your Draft and could 
> use the mecahisms you describe?
>
>I think the above sentence does not say that the mechanisms described
>later in the document can be applied to arbitrary syntaxes. It's just
>a general statement in the introduction to motivate that there are more
>than one way of matching, as described later in the draft.

I do not understand what you want to say here. A different syntax 
means a different ABNF. The "such" does not refer to mechanisms 
(there is no mechanism defined in the RFC 3066bis Draft), it refers 
to "language identifiers" what should mean ABNF plus registry.

But it is true that there should be several mechanisms defined (or 
simply exemplified) as some could be very specialised. Just one to 
give a practical example: I am a news agency and I survey RSS feeds 
to build a calendar. I am interested in pages where the date is 
expressed in a way my program and I can read it, among a list of 
languages I have someone more or less understand. I will therefore 
filter langtags in reference to the locale information on date 
format. This is a typical OPES job.

>A direct rewording that makes this clearer would be something like:
>
>"For any kind of language identifiers, including those defined by
>RFC3066bis, there is easily more than one of matching. Thus, this
>document defines a set of ways of matching for the language tags
>defined in RFC3066bis."

Full agreement. This clarifies the legitimacy of other types of 
language identifiers. ABNF + registry.

>However, I'm not sure it is necessary to start with that much generality,
>if the language tags of RFC3066bis need more than one way of matching,
>that's enough motivation.

IMO it really does because:
- it clarifies a text which I find - and probably many will - find confuse.
- I would then have no objection to the text.
- the wording could be slightly extended to underline what you point 
out: this kind of matching may make sense in other language tagging 
systems which uses the same subtags.

Regards.
jfc


_______________________________________________
Ltru mailing list
Ltru@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/ltru