W3C home > Mailing lists > Public > public-i18n-its@w3.org > July to September 2006

Re: Reply to last call comment on "Internationalization Tag Set"

From: Felix Sasaki <fsasaki@w3.org>
Date: Thu, 20 Jul 2006 00:48:50 +0900
Message-ID: <44BE5462.5040009@w3.org>
To: Steven Pemberton <steven.pemberton@cwi.nl>
Cc: w3c-html-wg@w3.org, public-i18n-its@w3.org
Hello Steven,

This is a personal reply.

Steven Pemberton wrote:
> Hello Felix,
> 
> Actually, your reply raised even more questions for us.

I guess we have a lot of misunderstandings. Would you / the HTML WG be
available for a joint call? That might make things easier to explain.
I'll try to make some points clearer below.

> Such as:
> 
>> See also the conformance section
>> http://www.w3.org/TR/2006/WD-its-20060518/#conformance-product-schema :
>> [[At least one of the following MUST be in the schema:
>>     * rules element
>>     * one of the local ITS attributes
>>     * span element
>>     * ruby element
> 
> XHTML 1.1 has both span and ruby. Does that make it conformant ITS?

as for ruby: yes. ruby is conform to ITS processing expectations, see
http://www.w3.org/TR/2006/WD-its-20060518/#conformance-product-processing-expectations
that is of course a cyclic statement, since ITS itself refers to the W3C
ruby TR.
However, that is in line with the general goal of ITS: to gather
existing standards for i18n / l10n purposes. In many other cases as well
we just point to existing work.

> How
> about XHTML 1.0 which only has span?

<span> is a piece of markup, which is not in the same namespace as
its:span, so it is not conform to ITS markup declarations, see
http://www.w3.org/TR/2006/WD-its-20060518/#conformance-product-schema .

> We don't understand how this could be conformant ITS.

A key concept in ITS is the differentiation between conformance related
to processing expectations and conformance related to markup
declarations. See the two links above. This differentiation is necessary
since we have a great variety of usage scenarios, see
http://www.w3.org/TR/2006/WD-its-20060518/#users-usage
That section makes clear that we would loose a lot of users if we would
not split up the conformance in markup versus processing expectations,
and separate them also for each data category. Now "separate them also
for each data category" means that a user of ITS can choose any
combination of ITS attributes  / elements - and will still be conform. I
think this is very different from the approach of XHTML modularization,
and you don't want to define a module for each possible combination.

> 
> But more specifically, if there is no normative schema, how can it (the
> spec) be tested?

First, the processing expectations can be tested. We are already
developing a test suite, which we hope to make public soon. And second,
the position of markup declarations can be tested. That of course has to
happen manually.

> 
>> - However, we want to follow the XHTML modularization approach, and we
>> would like to have an XHTML modularization example in our XML i18n BP
>> document, see http://www.w3.org/TR/2006/WD-xml-i18n-bp-20060518/
>> Also, in the XML i18n BP, we would recommend that any possible
>> proprietary modularization (as opposed to the ITS example, non-normative
>> one) related to HTML should follow the XHTML Modularization.
>>
>>
>> Would this address your concerns?
> 
> Modularisation allows you to add markup modules to a language so that
> you build a language using building bricks rather than defining elements
> and attributes separately. This helps consistency between related
> families of markup languages. We think that defining an ITS module would
> be a good way to help people get ITS into their markup languages.
> Modularization is schema-language neutral. Once you have defined a
> module, you can 'implement' it in any number of schema languages, such
> as DTD, XML Schema or Relax. We're not sure what you mean by
> 'proprietary' here.

I'm very sorry for this terminology and would like to take it back. The
point I (we) tried to make is that XHTML modularization and ITS just aim
on different levels:
- From my understanding, XHTML modularization is a framework for markup
modules which are defined in an abstract manner and can be implemented
in various schema languages.
- ITS describes data categories for i18n / l10n purposes which can be
used without a schema at all (see "processing expectations" above), and
from which a user can choose / implement whatever markup she wants. How
this can be tested is something we still have to prove in public, but
that will happen not later than CR.
The point of interrelation between XHTML mod and ITS is the "whatever
markup" above. I think XHTML modularization is a good way to implement
the ITS markup. My point is only that it is an optional choice to do so,
which is not related to conformance or any normative statements.

Best,

Felix


Received on Wednesday, 19 July 2006 15:49:05 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:43:07 UTC